USVI Public Records

Document text

k{ f ^ AUG 24 192^ NEW YORK ACADEMY OF SCIENCES SCIENTIFIC SURVEY OF Porto Rico and the Virgin Islands VOLUME IV—Part II Geology of the Virgin Islands, Culebra and Vieques Physiography, concluded—Howard A, Meyerhoff NEW YORK: Published by the Academy 1927 ' • ^'ir - f #.' NEW YORK ACADEMY OF SCIENCES SCIENTIFIC SURVEY OF Porto Rico and the Virgin Islands VOLUME IV—Part II Geology of the Virgin Islands, Culebra and Vieques Physiography, concluded—Howard A. Meyerhoff Ninv v(m:k : I'l I'.i.isHKn I'.v iin; A<-.\iu:mv ^, IIIK PIlYSKXxKAPHV OF THE \ IHCilX ISr.AXDS, CULEHHA AXD VIKQl ES By IIowaui) A. ^Ikvkuhoff ('()\TK\1'S Chapter lU The IMiysioi^raphy of S.iiiit Croix 11.1 (J('o^rai)hi<' outline 14."i riiysioLTraphic outline 147 The ohllaiid j \u liiH'k structures 150 Pliysioirraphie <le\-eloj)nieiit of the ol<llaii<l ir>-J The upland erosional surface ir»:{ Th(^ lo\v<'r erosi(»naI surface 157 The coastal plain 150 Tplift and wai-pini,' <ti' the coastal plain liV.] Dissection of the coastal |)lain 1()5 The suhniarine platform 170 The present level 175 Chapter I\' -The I'hysioi^raphy of \ icMpies and Cnlehi-a 1N| T!ie ohlland and its peneplanat ion 1st The lower i>enei)lane Iirj The coastal plain 11) [ Uplift and dissecti(»n of tlu' <'oastaI i>lain VM\ The suhinai'ine piatfoi'ni IPS Suhiner^^ence (d" the suhniarin<' jdatforni L*n7 The pi-esent level -Jll Uec(Md sliixht eniiM-LCence 'Jl:; Conclusions 2i:i Iiil>lio.i,M-a|>h\- 210 CIIAl"!'i:i{ III 1MIK IMn'SI()(;irVI>IlV OK SAINT ( hM)I\ (ii:()(d{ Ai'll K < )t I'M M Saint ('I'oix is tlie lari^cst of the \"ir^in Islands. It is nineteen miles lon^" from east to \v«'st, (ive mile- wide tliroui^h the western half of its leniith, narrowing- \(> less than three miles from Christ iansted eastward to Ta^uue I>ay, and to one mih' in the easternmost two and a hall* miles of its leniith. The hank upon which it rests is coiisiderahly larger, heini:- ( 145 I 146 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS thirty miles long by five to eight miles wide in its ontside dimensions (Plate III). The island is pushed so far into the northwestern corner of the bank that between Southwest Point and Ham Bluff on the w^est, and Ham Bluff and Salt Eiver on the norths the submarine platform is at no place so much as half a mile wide^ and in some localities it is wholly absent. On the north the sea-bottom deepens to 1000 fathoms within two or three miles of the island, and six miles north of Ham Bluff a sounding of 2412 fathoms is recorded. Westward, soundings are not available, but it is known that the bottom drops to 500 fathoms within three miles of the coast. Where Saint Croix narrows at Christiansted, the bank also contracts soutfiward, but not so much ; and from this point eastward the island is more centrally located. The platform north of the island is more than a mile wide from Christiansted eastward to Buck Island; there it bulges to the north again to a width of three miles. Beyond Saint Croix it narrows to a total width of four miles and extends east-northeastward for eight and one-half miles, where its bluntly rounded end rises sharply above depths of several hundred fathoms on the norths east and south. South of Saint Croix the bank varies between one and one-half and three miles in width. The surface of the platform is broken by three small islands and by many small banks and reefs (Fig. 37). Protestant Cay is a small limestone island immediately off Christiansted. Green Cay, situated three miles east-northeast of Christiansted, is a low island, 500 yards long by 150 yards wide, composed chiefly of diorite, intrusive into Upper Cretaceous sediments and volcanics. Buck Island, located three miles east-northeast of Green Cay, is one mile long by half a mile wide, and is composed of Upper Cretaceous eruptives or their derivatives. East of Saint Croix the platform is broken on the southeast by a breached line of drowned ridges six miles long, which rise 25 feet above its floor : and on the north and east it is broken by Lang Bank, six miles long and three-quarters of a mile wide, curving southward at its eastern end in a broad hook. Lang Bank rises 35 to 50 feet above the inner surface of the platform, and 70 to 100 feet above the outer marginal depths. Fringing and barrier reefs are abundant along the south shore of Saint Croix and along the north shore from Salt River eastward. The topography of Saint Croix is a decided contrast to that of any of the northern islands. It is divisible into three sharply contrasted portions: (1) a deeply dissected upland in the northwest, beginning one mile east of Frederiksted and extending northward to Ham Bluff and eastward to Salt River; (2) a wedge-shaped coastal plain (Part U Fig. 4), narrow in the west and broadening eastward until it crosses MEVHh'HOFF. /'f/r,x/Or,7,Mi'/ir OF Vll'fUIX li<LAMlS } |7 the islrtiuJ two miles iiortliwe^^i o( (/hristiMimted, in a ,«>rifs dI* disstH'ted i-uestas ; (3) a nmturdj disswtetl «r»«a of rugged liilk Itoiii ('hri>ti!iiistod to East Point (Fig. 3?). The tlirt'erohl division is ootirelv dopoodont upon the iiiidcrlyiiig rock>i. their Htrurtiire and pliysiogrit|)liie di'Voh>p- ment. Fiirtber description of tho topography will be presented in tlie iliseiissioo of tlie pliysiogrnphie formes. Physiogiiai'iiio Outlisk As ill the northern Virgin, Islands, the physiografdiie history of J^aint t'roix has as its l)!:U'Ji,groiind m\ oklhind of inassivi' tuej befU'led eruptives. with interhedded sfdinients, chiefly limestnius and stratified lull's. They w«>r.' more oomph'xly folded tlian tho .Hinular rooks of tlie saim- age in the iK.rthern islands, although it might h,> niur.' eorreet to >hv tliid: tdi.- • •nmphi'xity td' their folding is moro in lomhoieo. The rooks wero in^ truded hy dinrilos, but the fiitrusion diM»s not liavt,. the pxteiisivo devoh ..pno-nt that is found in Tortola. Hvvt Island and A'irgni (h>nla, or in wt'^toru \"iF<|U(S. Y»>t the small areas of deep-seated irdrusives luive produt-ed sntficiord^ offeid: u|mio tlie to])ograpliy to warrant individual nu'ution. Ihiriug tins stage (d' its history Saint Cr.uv apparentlv de- vchiped as part of a hirger mud. wldeh einhrae'orl the lau'thern Virgin h^lan.ls. fuh'hra, Mecpies and I'orto Hie-o. 148 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS Folding and uplift were followed by peneplanation^ which reduced the oldland to a rolling surface of low relief, now represented in the upland in the northwestern part of the island. The peneplane produced is probably represented in the crests of the rugged hills east of Chris- tiansted, but subsequent dissection has been so thorough that it may be- assumed, mainly by inference, that the erosional surface was formed there. Peneplanation was stopped before completion by uplift; a new cycle of erosion was initiated, and fluvial dissection of the upland produced a lowland of late mature topography 800 feet below the upper level. The lowland was submerged and upon it was deposited a series of Tertiary coastal-plain deposits ranging upward to 600 feet in thickness. Dif- ferential uplift followed, accompanied by warping of the upland and coastal plain and succeeded by dissection of both. As a result of the uplift, the coastal plain seems to have been emergent to the edge of the bank, and the fluvial denudation which followed was deepest and most thorough in th*e areas at the eastern end of the present platform. Although conclusive evidence cannot be adduced, it seems likely that the block-faulting which isolated Saint Croix from the islands north and northwest occurred approximately at this time. Depression followed, probably taking place in two stages, as in the northern Virgin Islands, the later one bringing the island approxi- mately to its present level and producing the features of submergence which characterize its shoreline today. Reef growth has proceeded vigorously; valleys have been partly silted up and barred; and^ on the whole, much of the shoreline exhibits considerable, but not complete, adjustment to the present level. Detailed consideration will be given to each of the physiographic stages in the order of their development, as follows : (1) The oldland. (2) The upland erosional surface. (3) The lower erosional surface. (4) The coastal plain. (5) The upraised and warped upland and coastal plain. (6) The dissected coastal plain. (a) The efnergent areas. (b) The submarine platform. (7) The present coastline of submergence. Wherever these events find counterparts in the physiographic history of the northern Virgin Islands, they will be briefly outlined and com- pared. Detailed consideration will be directed largely to the features which find no parallel in the northern group. MEYEKHOFF, PHYSIOGRAPHY OF } IROJy ISLANDS 140 Thp: Oldland The origin and development of the ohllands of Saint Croix and of the northern Virgin Islands were essentially alike. From the corre- spondence in rock types and the genetic relationship of the geologic structures in both areas, the conclusion seems warranted that during their early history they were parts of a single land mass. Definite cor- relation of the rocks in both areas is quite impos-jible, for it is doubtful whether any of the beds are exact equivalents. The agglomerates, tuffs and related eruptives of andesitic composition are indistinguishable from those in the northern group. Between unfossiliferous rocks of this nature specific correlation from one locality to another is impossible, especially when lateral variation and lensing are common features. Among the Cretaceous sedimentary rocks of Saint Croix there are several limestones, some of which are fossiliferous. The small fauna yielded no species or genera which were common to the limestones of Saint Thomas, yet the assemblage is equally characteristic of the Upper Cretaceous. Although the rocks of both island groups are of the same age and were accumulated and later deformed in the same manner, they may not be strictly contemporaneous and equivalent. Contributory evidence of the correspondence in their history is found in the nature of the deep-seated intrusives. The intrusives of Saint Croix have a restricted occurrence, attaining importance only in two localities—in the Green Cay-Sonthgate region, three miles east of Christiansted, and in the Fountain basin, one and one-half miles southeast of Mount Eagle. In both localities rapid decomposition and disintegration have reduced them to lowlands. From their mineral composition, texture and relation to the surface volcanics, their genesis, it may be inferred, paralleled that of the intrusives of Vieques and the northern islands. Except for the areas underlain by the intrusives, which have been easily eroded into lowlands, the various rock types composing the old- land have exerted but little influence upon the topography. They are divisible into the three types described for the northern Virgin Islands, but the erosional forms of the thinly stratified volcanics and sediments are the same as those produced on the massive eruptives and shallow intrusives. At Watch Ho (Vagthus Point) on the south shore of the island, three miles south of Christiansted, a series of vertical limestone beds have produced a vertical escarpment coincident with the bedding planes for a short distance along the coast; but even here, upon the summit of the promontory formed by the limestone, the effect of the thin bedding upon the drainage is lost. Other comparable examples of dip-slopes and escarpments along the shoreline upon the rocks of the 150 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS oldland might be mentioned; but, on the whole, shorelines, drainage lines and ridges exhibit a supreme indifference to rock types. EOCK STRUCTURES The structure of the oldland was carefully studied and plotted by John T. Quin, F. R. G. S., long a resident of Christiansted and an enthusiastic student of some phases of the geology of Saint Croix. His study, although intermittent and undertaken only as a hobby, was ex- tended over a long period of years and is based upon much more exten- sive observations than could be made in the relatively brief time which the writer had to survey the island. The present survey, therefore, affords but partial confirmation of Dr. Quiri's observations throughout much of the island and supplements them elsewhere. In the main Dr. Quin's results, given in his work, "The Building of an Island,^'^* will be used in outlining the geological structure. At the present time the oldland is represented in two large areas, subequal in size and separated by the thin prong of the Tertiary coastal plain which crosses the island one and one-half miles northwest of Christiansted. One block occupies the northwestern part of the island : the other, offset to the south, comprises the entire eastern half and includes Green Cay and Buck Island (Fig. 37). In both areas tlie geologic structure seems to be a comparatively simple system of folds, crowded in the -eastern block, more w^idely spaced in the western. In the western section the dips indicate the presence of a single anticline and syncline, which |)lunge southeast, and w^hose axes trend N. 55^-60° W. The syncline cuts diagonally across the center of the block, and the anticline clips the southwestern quarter. Crest and trough are approxi- matfdy three miles apart. Xo indications of other folds have been found, and the prevalently massive nature of the rocks makes the possibility of other large structures remote. Some minor faulting has taken place, but in no observed case was the throw greater than 100 feet. Con- tortion of strata with minor crumpling was noted, particularly at Ham Bluff. In the eastern half of Saint Croix three anticlines and syn- clines have been plotted by Dr. Quin. If his interpretations are correct, the folds are asynmietrical, with steep to vertical dips upon their south- western limbs. The crests and troughs, whose axes strike N". 60°-75° W., are from one-half to three miles apart. They pitch southeast, but the pitch is not strong, as in the folds of the northwestern block. 2* QuiB, J. T. : '*The BuUdIng of an Island, being a sketch of the geological structure of the Danish West Indian Island of Saint Croix, or Santa Cruz." Published by the author In Christiansted, Saint Croix, 1907. Map (frontispiece), and Chapter V, pp. 44-52. MKYKUHilFF. I'HYsHM; IIM'II V OF IIIHSIN l^sLAM)S 131 Th(;*r' folds ^.xert « negligible elfVi-t 011 tlie toiM.m-ai.hy, altliough, lortillY, H'liittll ser-tioriH of tli« eortrtiine are di'terniiried by Uif strike or the T<K'kH; fur example, at Watrb Mo nn the suutb-ceiitnd eoa^t. The Htrike <'riidely parallek the eoa^^t fnirn Fnis|>erity (o Salt River Point on the iiorrli sliore. yer does not uppertr to mold it< fnrin. Tlie north- westerly <lirecli(>n of the eoast frt)Oi Christ iansted to dudiths Fanev conforms with the struetiiral trend, but tdher j;oi<lt»nee irn|dits thrtt the outline of this j>Hrt (d* the eoast wa.- formed without r»dVrenee tn tlie Htrneture, The valley from (ireat Pnnd through Luwry Hill to Altona .Lagoon is the only nmjor feature in the to[iograpby tlie direc-rioii of w^ldeh seems to he a direct eonsi'queuee of the trend of folding. It et.ineides with the erest of an anticline. Imt it is to he confessed, that lichl l"'Hi. :iH. Tin- ftlltif liiirhttlfl, flU, ohsio'vattons lM,>re md: the io,tcr|iretation only in tin* Fort Louisa Augunia |>euiusnla. where the crest of the anti«'linc is not excavated into a vaHey as it is to the sontJu-ast. In phort. tlicre is not a single important topo- graphic feature the; devclopnicid: of whi«-h i-an lie referred with rcrtaird^v to the underlying iVdda. ilirior norjual faults are nnmerou.s. hut. -o far as oould be obs<«rved.. none of ilieni has topngra(dd<' (expression. From a point a udle ofTshorc the writer rauglit a suggestion of stcp-faultiug and tdtiug in tlie ruggr><l lijlls constituting tlu^ easterunu>st three udk's of the island. Several suceessivr> hills, carved in complex struetiircs, pn«seutc<l slwp western fronts and gcnflnr eastern haek^slopcs. As the imprr-ssi«m was olitained eidy from one pmnt of! the imrth coast ami was not e,inlirm«"d, no emdi- dent'c is placed in tln> opinion. Judging from lH\ C^lnin's observations, it swans equally possdih- that flie hilL are lmgl)ack>. 152 ^SCIENTIFIC ISVRVEY OF THE VIRGIN ISLANDS The breach between the two areas of oldland rocks^ filled by Tertiary limestones^ appears to have been excavated on a major fault- whose strike lies between north-northeast and northeast (Fig. 38). The evidence for faulting rests upon the following points : (1) The presence of the only lowland which -crosses the island, and which is now occupied by coastal-plain deposits whose accumulation it antedated. (2) The structural discontinuity of the two blocks. If tlie anticlinal axes are projected from the eastern to the western areas, they meet no corresponding structures. (3) The presence of limestone beds carrying the fossil Eudistes at Watch Ho on the south, and similar, but thinner, limestone strata at Judiths Fancy on the north. If these two limestones are assumed to be equivalent, their position indicates an offset of the w^estern block to the north, of the eastern block to the south. If they are not equivalent, tlie presence or absence of corresponding limestones in the two blocks is not significant, because of the southeastward pitch of the structures. (4) The southward flexure of the strike in the beds at Judiths Fancy. This feature may be drag caused by horizontal movement of the western block to the north. (5) The presence of a broad coastal plain overlapping the southern margin of the western block upon a well-defined lowland ; absence of both a distinct lowland and Tertiary deposits on the seaward margin of the eastern block: The faulting appears to have occurred prior to the development of the lower erosional surface upon which the coastal plain rests. Save for this single line of w^eakness and offset, faults are not important as a cause of any topographic features of the present oldland areas. It is to be remembered, however, that the boundaries on all sides of Saint (Voix bank are block-fault boundaries, those on the west and north, at least, being of comparatively late origin. PltYSlQOUAPHIO DEVKLOFMENT OF THE OLDLAXD The similarity of the Upper Cretaceous rocks composing Saint Croix to those making up the northern islands, in age, origin, lithology, structure and mutual relationships, is conclusive evidence that both regions were initially parts of the same land mass. The clear-cut fault boundaries north of Saint Croix and south of the northern Virgin Islands indicate profound block-faulting, which has formed between them a graben some 10,000 to 16,000 feet deep and but 24 mifcs wide. UI-:¥l-:h'UOFF. l'IlY?<hH]l{ \PMY OF Mh'JUS /KLJ.V/IK ^ :,:| Tlu' |iusiti(Hi u( Saint Crtiix ill thi' Mo^^ozuit: irtiiil iiirt>s iM-forc ulork-faiiltiu*; J!^ by W h et h (' r i t w a k i-h-ivIy H SOUtil.TJl projfctioii from lh»» iiorthorn i>ilan«ls or u j.iirt of a .siihstaiitial iiia!^.< i-i»iitinii()iiK wilh l'<>rtu JVu-o i-ninwi ho. i|<'teniiiiH><L Wliai- .v«-r itH coimc<-Tioii8. <i.'|M,sitiou, luldin^tr i and jiiji'crioii of ilw I I ru(-ks, and pprsistod ^ | iliroiijrh the poTKMM | i whlvh followed. Ill = all rosiMH-te the early t history of tin. old la nil | <tf >Siiiiit I'roix was | similar to that of tln» ^ ! noHlHnni Virgin Is- ^ | hiiKis oldknd. 'riiK rd'TiAxn Kiio- SIOXAL Kl'liFACl.; From <«<iv('rnnu'nl Hill. Saint Thoina,-, tho islam! of Saint Cruix may he mh'ii <,u rh-Ar days onier-in- partly a Ihm <• t hv itnithorn horizon. In flM« ooiirf'i: fd' till' is- iiim! ,01 tlu' v,(-i„iiv of f'liri<1ianst<.d, J 5 1 f<vii-:xT/Fi(' i>i(:ii\i:'y (fF the viRciy islaxdi^ there is ii bvemi} gap. Huec;ee<Jeil westward by a .series (if low rolling hills, a gap iiuirking the position of Halt I{i\x-r. then a series of higiier lulls whicli rise and full (.'onsitlerablv, but which at no point <Jrop bt>h)W ibe borizniL 'Fhe Jiccordanoo of llio tops of the bills in western Saint Croix is a nutieo nble iVatiin.. luist of the fhristiaiisted g«p the hills arc- rugged: yet there, too, chs.-^pite iioteln's whieli pnss ht'iiejith the liorizoii, the tendeiioy is toward elevations whieh neeord with the ludgbt of the western npbiiHl Aotual slndy of the island eorrohorafes eonvineingly the (jistarit 1ni- prtjssionsi obtained from Saini Tlioinas. The old land which eonstitute.s n{»rthwe«tern Saint C'roix rise.s to a ileeply dis8e<;t.ed upland, which,, prior to difjsoetion, was a ^ rolling surface of hills in an erfisional stage <jf early old age (Fig. ;I9). The Imselevel appan-ntly wa;^ 150 to HfHi fccf above I he present position o( the sea, and above it rose irregular hills with, 100 to ;MT, feet of rdivf. The most prononnecd hill gronp is the Mount l^]a«1e-BhH! ,Monid.ain mass, whoso in,a,xirniun olevis^ tion is lli;:. feet rtfwve 8cad,evel. Bodki,n,, two ndlcs west of ,Mouiit, Kagle. is WKl feet high, and the other summits range between H'^o anil !»'i.") IVct in height. From any ennnenee or ridge upon the upland, its once sm<M>thly numded slopes, chararteristic of old age. are more con- \incing than when tlun' are viewtnl from, a (listance or subjected io ton minute analysis of topograpbic detail (Fig. 40). I'he <1,eeply incise.f. Mim-iiuiOFF. f\fii\siof!M.U'ny OF viMaix i.<LAMhs ^i:,:> v;i!lin-s ri<i\r disswtitig- its uiNlu'Iiiliiig^ siirfrtn' »n; uttiMiv i1i«-<>r«lrtiit with its hijImIu...! mi.l rnyiidiHl Umns | Fin-. 41,. The iiphunl siirfaeo riurrks the tiihiiinatioii of ihe earlit^Hl rvrunhMl unmtm vych. It; appears^ to be eqiiivuhnil t:u the Sn'mt John poiieplane I'f the ixirthem ishirids, hut tlii;^ i-onelusioji is open 1o question. In •rt'iieral it Iws not attaiiiefl thi- degree of perfeetion whieli tlie p.-ne- «.(•<;-. eHl-(;v<de leve], an ropre.^entfnl in eiistern Saint Thonnis (vf. Viiis. .')!» laid Fiiii I. -iO). h\ eh'vatioii it is suinewhat hi|i;her than ihe larler, hnt it i> hnrer than the Saint dtdui |M'neplane. Then' \k niore.nnr. no intia-- inediare erosion, snrJaH;e in Saint i'r^nK hetweiai the ripL-ind and tlie lower penep'lane npon whieh the Tertiary eoastal phdn, was depositenh :rher<' arc tlms reeorded hnt two ereision e-vides, as e<iinpared with three ill tile nortlun-n iskiids. On the Initpr the Tertiary eoastal phiin is f-ntirely siilHner<>'ed, but in S^int Croix it attains a inaxinjinn eU'vation i.r WH) feet above nea-hweh If tlie eomparalive ek-vations of tlie r<iastal plains are <-onsidered to ri'present aj)proxinnjtelv the anuainl of dilfenai- tial uplift between the nortliern jsIhihIh and KaJiit Croix, the nphual. h.'^mning hnt litth' nn^re than 2m) h-et above the highest eoasfahptain l.efls. would eorre>])ond to the secianCeveh' <n-osi«aud s„rfi,ee. whi.-h lies aianit 'MM fcot aisove the subnn-n/od ooastnl oliuhj in th.- noridu.rn Vir,..o-n i:,ij sr/f:xTii-i(' sf;ini':r of thh iina/x i.^i.amu l»i w figures ('heck \i^'1t eluscly. On \'ie«|iies, when: the r).n.<t!il plain attains an olevation uf 150 [oef. tlie lirst-cyr-lo p.^n.- pliitic is absent and the sceorid lies at a,n <'l(>vatHHi ul JjMt feet. Tlie fcIatJen l>etween the I'ertiary and seeend-eycle penephiiie there is t:lii!> the <ain<' as tlie rehitiun 'between the <t,axtal plain and npiand in Sahit: Croix. On t!ie otJn-r haial, it nniy l)e poitda-d unt thai flie Saint Croix npland aijprnxinnites a peneplane mere nearly than the upland surface of Ti>r^ tola or Saint 11 id that for this reason th(- dem-ee of Fes. 12. Tiir Inwrr w fhmloielr fro^lnuitj .-iin-fm e. flankitiff Iftf tjin^c-tnl upland iud:ii»n. as eotnpiired with tlie Saint Jolm |>eneplane, is not a valid m- terion : that Ihe eoastal plain, thinker than in the northern islands, nuiv have hnried the seeornl-eyide nnrfaee: or the lattnr way have la-eii e«an- pletely dr.«troy('d during later disse«-tfon to a lower haselevel. 'Hi., weight of argument, however, favors the oonelusioii that th»' nplanri ..f Sainr Cr.nx eornspcauls to the seeond-eyele siirfaee of the northern ishmds. and it will he eonsiilered hereafter to have that relation. 'Flie (dillaml hilltops of eastern Saint (,'roix apprnxinnite the miniinuni oleYationfj of the western uplainl and may he eoiifiidered as a remnant wideh has sinee hi-i-n \ery strongly dis<e«dvd. meyi:rh()ff, physiography of viRiUK jslaxds ir>7 TiTE Lower Ekosioxal SruFACE Peneplanation of the present upland was st()])pe(l by uplift of 700 to 800 feet, an amount which corresponds exactly to that which occurred between the second and third erosion cycles in the nortliern Virgin Islands. Denudation to the new baselevel proceeded until a rolling to rugged surface was formed (Fig. 42) and until only two small sections remained in which the earlier erosional level was ])reserved. Miu'h of the lower surface formed is now buried beiieatli the Tertiary coastal plain; hence its nature cannot be accurately ascertained. The valleys cut into the upland represent no more than slender extensions which are not typical of the entire surface, and even the inner lowlaml areas, where the coastal plain has been stripped, are marginal to the original surface and can scarcely be considered typical. The margiiuil sections exposed locally in the inner lowland, where the Tertiary marls and linu^- stones have been, removed, are hilly and alnu)st rugged. Their char- acter is well displayed between Saint Jolui estate, three miles west of Christiansted, westward to Little Fountain. Here contacts between the Cretaceous and the overlapping Tertiary, although usually obscured by surface wash, are not uncommon, and the boundary can everywhere be drawn with a fair measure of accuracy. The contact is found through a wide range of elevations, differing in short distances l)y so much as 150 feet. West of Little Fountain the marls and limestones abut directly against the escarpment between the upland and the lower ero- sional surface, burying the latter almost com[)letely. East of Christiansted much of the dissection of the Cretaceous upland was accomplished during this stage of physiographic history. More recent erosion has been engaged largely in removing the flanking coastal plain, a single patch of which still remains on the hill slope above Milord Point, while a few scattered Tertiary fossils were found half a mile or more inland at two other localities fronting the south coast. The coastal plain was, therefore, deposited along the margins of the eastern oldland block, and the surface upon which it rested was a rugged one of fluvial dissection, corresponding in age and development to the one underlying the Tertiary to the west. North of Saint Croix, Green Cay and Buck Lsland rise as monadnocks above the lower erosional surface and above wliatever may remain of the coastal l^lain sulimerged beneath the mar- ginal waters upon the platform. The second erosion cycle recorded on Saint Croix is tluis found in the surface upon which the overlapping Tertiary rests. In its relation to the Tertiary it is similar to Lobeck's lower peiu^])lane on Porto Pico, and to the third-cycle peneplane of the northern Virgin Islands, to both l.'l SilESTIFIV sriiVHY <tF THE VlRdlX IsLMXHS of wliieh It is (N|iiivijleiit ,iii nge and tiiudc of ori,Kiiu Its niw-ediU;W8 m the (uitcdiin- (>r liuviiil (li^s.'<-!ioii. which prodiujed ii Jitt»' iiiatiLrc sih'IVh-f i>l' Um to rii«.d<-ratr'-rp|ier on tin- wrtward portion o( tlw biiili, nhovc wiiirh a few residuals rise. >iK-h as liiirk Ishiiid and Uiwii Cay: iiiatuiH!, nig-god t«»|)o.ui'a|ihy ..f miidcratc lo hi-^h relief in the intcrifir, mvh ha tli«' i'liHtcrn oldland aiva and th«' rtaithcrn and wiMwii hin-der.s of tJic iiortliwcsfcni uphmd: and tlio sidnnalnre to jnatniv <li>!^tH;ti()n ot: !lu> roprf'Hcntativi'. Ihiriii.j'^ thf dis^:^•.•ti«m r.f th,. nnland sevoml portion^^ of the Mim<\ line lw>twei'!i the northwestern mid eastern, uldland r'enuiajd\s ( Fiy. .18) were ihe h»r«i-e>t areah redni-«'d to loudands. I)iit: several others deserve inoiition. Between Fonntjun and JJiver a ileep hasinal vaHey over half a, niih' wi<le was ertided out ( Fi,j£. i^h in large part hwniwe of tin; presence of a dicnate intrnsion in the upper part of the vallev. South of Munnt Ka-ie, ifoinediately east ..f the Fonntain-Iliver vaUey, is a, .sniiill hasin with an onrlet to the sonth and a snudi pas> to tlie snutli-^ east. Xo featinv ront rolling its developnirtil eonhl he deteeted,. The hrortti flopression from ^?onth,^ate• J'ojid aeros.s the islainl t<» (Irv-.n Pond was developed partly on cpriekly disintegratetl and dee(anpos«.d iliorite MEYERHOFF. PHYi^lOGRAPB Y (IF \ [RdlX ISLAXDS 159 on the north, but on the south alluvial jnaterial lias eoni])letely masked the iin(lerlyin«' rocks. The rocks flankino- the depression are all of vol- canic derivation. All of the other valleys in the upland are of normal size and development. Under ordinary conditions of fluvial ])ene])la nation with equal distri- bution of rainfall on homogeneous materials, denudation is greatest at the coast and becomes progressively less inland. This relation holds for the southern, eastern and northeastern portions of Saint Croix. The northwestern upland, therefore^ requires e>;|)lanation as an abnornuil feature. It was not completely reduced during the second cycle (of Saint Croix) ; yet the least reduced areas, from Mount Eagle to Maroon Ividge, constitute the northern coast of tlie island, notwithstanding the fact that they receive a heavy rainfall, and that their rocks are not superior in resistance. The margin of this part of the platform is a block-fault boundary. The faultiiig wdiich formed it must have been of such recent date that scarcely any dissection of the coastline has occurred (Fig. 43). There is little or no submarine platform, no coastal plain, and no adjustment of drainage to tlie ncAv coast. The streams dissect- ing the upland are more or less adjusted to gradients accordant with the contact between coastal plain and oldland ; and although many of them head within a mile of the northern coast, they flow away from it toward the south. Absence of the lower erosional surface off northwestern Saint Croix, together with the other factors considered, forces the con- clusion that these features are not absent because tliey never developed, for they must have developed. Their absence can be explained only as a result of comparatively recent faulting, whose date is tentatively placed as Pliocene. Erosion of the lower surface proceeded no farther than late maturity on any part of the oldland, because the cycle was terminated by Middle Tertiary submergence, during which deposition of the coastal plain took place. Drowning seems to have exerted l)ut little effect upon the rugged surface over which the sea advanced, for there is no evidence of wave planation or other marine modifications except deposition. The Coastal Plaix Resting upon the erosional surface just descrif)ed is a series of gently dipping coastal-plain depcsits, consisting of a l)asal calcareous con- glomerate, and alternating foramini feral marls, firmly consolidated cal- carenytes and fossiliferous crystalline limestcuies. They vary greatly in total thickness, but the variation is primarily the result of subsequent erosion and is only in part due to thf> irregularity of the surface upon 160 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS which they were formed. Their maximum thickness is displayed in the series of outcrops from Golden Kock estate on the road between Chris- tiansted and Judiths Fancy to the crest of Bulowminde. Golden Eock ib 75 feet above sea-level and lies but little higher than the average elevation of the contact between the base of the coastal plain and the Cretaceous rocks which they overlap. Bulowminde is located on the top of a cuesta in which the coastal plain attains its greatest elevation, 590 feet. If allowance is made for erosion of the cuesta, the original coastal plain was not less than 550 feet thick, and was probably nearer 600. The fauna whicb was collected from the strata has not been studied ; hence a final opinion, of their age must be deferred to a later chapter of this report. Yaughan assigns them to three different geo- logical horizons—Middle and Upper Oligocene and Lower Miocene. -'"^ Physiographic evidence impels the writer to date the entire formation tentatively as Oligocene. Its age, however, is of no moment in the present discussion and will be referred to noncommittally as Middle Tertiary, but in correlation the formations will be tentatively assigned to the late Oligocene. Two lithologic features of the coastal-plain materials are of physio- graphic significance: the thin basal conglomerate containing fragments of oldland rocks, and the absence of oldland material in all the beds above the basal conglomerate. The conglomerate is the normal product of deposition by an encroaching sea upon a rugged fluvial surface; and, as might be expected, it is found only in the lower or valley portions of the contact, and is normally absent from the hills which the coastal plain buried. The nature and distribution of the bed indicate how slightly the underlying surface was modified during submergence and suggest that burial of the surface was rapid. The absence of recog- nizable fragments of oldland rocks from the balance of the coastal-plain deposits and the presence of but small amounts of clay derived from decomposition of the Cretaceous quartz-free volcanics indicate that com- paratively little erosion of the oldland occurred during the formation of the coastal plain, and that the rate of limestone deposition was much more rapid than the rate of erosion in the adjacent land areas. During the final stages of deposition, it is to be remembered, sea-level was within 225 feet of the upland erosion level, and extensive erosion could not be expected. At the present time, despite the elevation of the Virgin Islands above sea-level, little land waste is being transported to the submarine platform areas to add to the accumulating coastal plain. ^ Vaiighan, T. W. : "Stratigraphy of the Virgin Islands of the United States and of Culobra and Vieques Islands, and Notes on Eastern Porto Rico." Jour. Wash. Acad Sci., Vol. XIII, 1923. p. 311. MEYERHOFF, PHYSIOGRAPHY OF VIRGIN ISLANDS 161 Even on rugged shorelines where beaches are present^ coral-sand accumu- lation exceeds rock-debris accumulation, and within ten yards of a head- land coral-sand completely displaces rock fragments in the beach ma- terial. The only important deposits of rock waste of recent origin are fluvial and estuarine. Apparently the conditions of Middle Tertiary coastal-plain formation were analogous. The Tertiary coastal plain of Saint Croix flanks the northwestern oldland in a narrow strip *of increasing width from William estate, one and one-quarter miles north of Frederiksted, southward to Concordia estate. There the line of contact swings northeastward, and along the south side of the island the belt of Tertiary gradually broadens as the oldland narrows (Plate III). Starting with a width of one and one- half miles, it increases to four miles at Windsor and Morningstar estates, and immediately west of Christiansted it crosses the entire island. South of Christiansted it ends against the uneven surface of the oldland in an irregular series of cuestas (Fig. 38). In the lowland south of Catherine's Rest it has yielded place to alluvial material mixed with Tertiary limestone fragments; on the south shore it retreats westward around a low promontory of Cretaceous rocks which make the cliffs at Watch Ho. In the eastern half of the island coastal-plain deposits were found in situ at only one point—a small patch above Milord Point, on the south coast; but a few weathered Tertiary corals were picked up in the alluvium in two of the larger valleys fronting the Caribbean (north of Longford estate and in Cottongrove valley). On the northern side of Saint Croix the coastal plain is found at a single point,—an outlier of the southern belt, which extends completely across the island for a short distance on either side of Little Princess estate, northwest of Christiansted. Eastward no coastal-plain material was found, unless the limestone composing Protestant Cay should prove to be Tertiary. It is assumed to be Pleistocene or Recent because of its position within the present reefs, but it must be admitted in the absence of paleontologic evidence that Protestant Cay may be a remnant of the Tertiary which has survived later dissection. Along the northwestern upland there is neither a coastal plain nor a lowland upon which one might have rested. Since its deposition the Tertiary has been deeply dissected and a con- siderable portion of it has been removed. Its present restriction to the south side of the island provokes an inquiry concerning its original extent. That it w^^ developed on the northern flank of the northwestern oldland prior to faulting is a reasonable conclusion, but the outlines of the coast in that direction before the present fault-coast was formed are not known. East of Christiansted, however, neither the north nor 16^ SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS the south coast has been directly affected by faulting since Middle Terti- ary time. The relation of the coastal plain to this part of Saint Croix is, therefore, a legitimate field for inquiry. It was pointed out above that the oldland in northwestern Saint Croix was severed from the oldland in the east by faulting, which occurred before the formation of the lower erosion level. Erosion carved it into a lowland, and during Middle Tertiary it was filled with coastal-plain deposits (Fig. 38). The ques- tion arises whether later faulting along this line of weakness could have occurred, thus causing the paucity of Tertiary deposits in the eastern section of the island. Field evidence is decidedly negative. Although the coastal plain fills the faulted area and peters out rapidly east of it, the strata occur on both sides of the fault-line, while Cretaceous rocks rise as monadnocks through the Tertiary west of the line. The coastal plain has the appearance of a deposit formed in an embayment between promontories and is without specific relation to the fault. Moreover, the occurrence of Tertiary drift and of one small patch in situ on the southern slopes of the eastern oldland area is sufficient evidence to justify the assumption that it was formerly present along the entire southern coast of Saint Croix. No field evidence was obtained at any point along the northern coast east of Christiansted favoring the view that the Tertiary was deposited marginal to the oldland. It is hoped that more detailed search will reveal residual patches and float that will prove its former existence. Even without field verification, the conclusion that the coastal plain was deposited there is inescapable. Foundation for the conclusion rests on the following points: (1) The presence of the mature erosional surface, eqiiivalent to that upon which the Tertiary was deposited on the south side of the island (Fig. 37). (2) The extension of the lower coastal-plain beds to Little Princess estate and possibly to Protestant Cay, a fact which makes it very likely that the upper beds, at least, extended much farther along the north coast. (3) The total thickness of the coastal plain, which is such that it must have occupied every portion of the lower erosional surface. The removal of the coastal plain from this area need not be con- sidered as complete; for, as will be shown when the submarine plat- form is discussed, some coastal-plain material still exists submerged. The removal of so much of it that none remains upon the eastern section of the island is not surprising when it is considered how narrow the eastern width of the entire Saint Croix bank is as compared with the southwestern. Thus in the 'first and second erosion cycles the MEYERHOFF, PHYSIOGRAPHY OF VIRGIN ISLANDS 163 eastern end of the island was thoroughly dissected^ while in the western portion dissection was imperfect. Similarly, when the coastal plain emerged, dissection and removal were more thoroughly accomplished in the east, because the streams working headward from the north and south coasts had relatively short distances to go to establish gradients to the drainage divide. The coastal plain is believed, therefore, to have extended completely around the island. Its width was variable and depended upon the width of the lowland developed prior to Tertiary submergence. AYhere this was narrowest, subsequent erosion has since largely or completely re- moved the deposits. Along the northern coast of the northwestern old- land area the coastal plain has been faulted into the graben between Saint Croix and the northern Virgin Islands. Uplift and Warping of the Coastal Plain The deposition of the coastal plain probably occurred in the Middle and Upper Oligocene, as on Porto Eico. No erosional break could be detected in the series, and they may be regarded, therefore, as a unit of accumulation. Following their formation, they were elevated at least 600 feet; but the movement was slightly differential and the amount of uplift differs from place to place by small elevations. The base of the Tertiary at the foot of the northwestern oldland lies at heights varying between 100 ^and 200 feet above sea-level. Elevations range but little above 100 feet from Wheel of Fortune estate east- northeastward to Fredensfeld, but at Windsor and locally in the vicinity of Dolby Hill the contact between Cretaceous and Tertiary reaches 200 feet. At Little Princess estate, on the other hand, the contact passes below sea-level. The irregularities are caused by the unevenness of the surface upon which deposition occurred, or have resulted from the removal of the coastal plain to varying distances from the oldland. None of them can be definitely traced to warping, which played too small a part in uplift to affect seriously the average level of the contact. Vaughan has identified the Tertiary near Wheel of Fortune as Middle Oligocene; that along the north shore west of Christiansted, as "prob- ably Upper Oligocene." Physiographic evidence is at variance with this stratigraphic determination. The Tertiary west of Christiansted lies at a lower elevation than the beds at Wheel of Fortune, and the uplift of the coastal plain was not accompanied by any folding or warping suffi- ciently strong to Justify the conclusion that Upper Oligocene beds on the east are at a lower elevation than Middle Oligocene strata on the west.^^ '^ stratigraphic and pnleontologic evidence of the relative ages of the^ beds will be considered in an ensuing part of the report. 164 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS The uplift of the coastal plain was, however, accompanied by slight warping in a direction the reverse of that implied by Vaughan^s strati- graphic determinations. The deformation produced a broad, shallow, semicircular structural basin between Kingshill and Beeston Hill. At no point were the strata observed to dip so much as five degrees. The principal movement was low arching along a northwest axis frOm Chris- tiansted to Salt River Point, with a subordinate axis of arching from Fair Plain northward to Windsor estate. Erosion on the crests of these arches has produced cuestas, which form a semicircle around the struc- tural basin between the two axes of uplift. Other cuesta development has had little or no reference to warping or folding, but is the result chiefly of the initial seaward dip of the strata. Quin has reported westward dips up to 15° in the Tertiary along the west coast of Saint Croix. In two days of observations along the shore- line from Ham Bluff to West End Salt Pond the writer was unable to discover any dips of that magnitude. The strata dip consistently toward the west, but the amount is less than five degrees. The structure is apparently original and normal in deposits laid upon a seaward- sloping surface. It was hoped that steep dips formed by drag in block- faulting would be found and thus substantiate the belief that faulting has occurred at the west end of the island since coastal-plain accumu- lation. Dr. Quin'^s more extended observations may have brought to light a locality overlooked by the writer, but certainly the feature has a very limited development, if any. With uplift of the coastal plain a few other minor flexures were formed in the Tertiary beds, including low arching in the Bethlehem- Fair Plain valley and in the Grove Place-Diamond-Enfield Green val- ley. It is scarcely detectable in the structures of the beds, but was sufficient for the development of shallow subsequent valleys flanked by cuestas. According to Quin there is structural evidence for arching along a northeaSt-southwest axis from Saint George Hill, at the south- western corner of the upland, to Southwest Point. This arch has no physiographic expression, unless it explains the Southwest-Sandy points peninsula. The structural undulations imparted to the Tertiary were shared by the oldland; but, because of the mature dissection of the upland in which drainage lines were well established, the low arching and depression made little determinable change upon its surface. Two features in the physiography of the oldland may, however, be the out- come of the post-Oligocene uplift. Between Mon Bijou and Little Fountain the coastal-plain limestones terminate at, or possibly are buried MEYERHOFF, PHYSIOGRAPHY OF VIRGIN ISLANDS 165 beneath, an alluvial fan made up of Cretaceous rock fragments which were washed out of Canaan valley. Gravels from the same source were also found at Glynn estate. The Bethlehem arch seems to have caused more rapid erosion in Canaan valley, and to have resulted eventually in deflection of the Canaan drainage into Salt Eiver. On the west flank of this arch the drainage of Hermitage basin appears also to have been diverted from the Colquhoun-Mt. Pleasant valley through the low pass to the southwest. The arch detected by Dr. Quin through Saint George Hill may have deflected the Annaly-Mt. Stewart-Montpellier drainage from the Two Friends-Grove Place route to the southwest past Orange Grove. The stream capture which occurred here may have been facili- tated by the uplift. Dissection of the Coastj^l Plain Erosion has worked upon different parts of the elevated coastal plain at such dissimilar rates of speed that the western portion has been dis- sected to an old plain; the central portion, to a late mature surface of low relief; the eastern portion southwest of Christiansted, to a mature surface of moderate relief. These three divisions, although distinct, are transitional, the mature forms passing imperceptibly into the old from east to west. Each division merits brief mention. The Tertiary coastal plain west of Diamond and Enfield Green estates, as far as the Saint George arch, is well drained by intermittent consequent streams originating upon it or flowing southward from the oldland. The Tertiary west of the arch is poorly drained to the west or northwest. Much of this area has the aspect of a very young coastal plain upon which drainage lines have not yet been established. From a surficial examination of the topography, it was difficult not to assume that the southwest part of the island was made of younger materials which had but recently been uplifted above the sea. Insufficient paleon- tological work has been done up to the present time to enable the writer to have a first-hand opinion on this point; but, judging from the conclusions of other workers and from every line of field evidence, the Tertiary in the west is not younger than that east of the Diamond- Enfield Green valley. If anything, it is somewhat older, for the strata are continuous with the lower and older Tertiary strata in the coastal plain farther east. Removal of the higher resistant layers, which are still present in the east, has exposed soft marly limestones, whose uniform non-resistance to erosion makes the development of strong drainage lines impossible. Structural undulations have thus produced no decisive topographic effect, and the surface is a smooth seaward- :U)Ci sCir.STlFii- slllWEY OF THK VIRaiN ISLAMJS Hl(^f>iIl.^^ plain which pvvs liuJe evi<lcrice of c]e;»p disH»;n;ti<iii. A.^ wlitary exceptions tu till' iiiil)roki']i .-lope, the uHHtxTii iiiardii of the town of l''iv(|erik.<t('(l fill!-, in a Itoic.stono ofarpnicnt 10 to 50 feot hig-h. fneni^ inlan.l upon n iinrrow lowhuiil (Fig-. U), mul the nmd fmni Ln lirango soiithwiird past \\1hh'1 of Fortune to I'ojieordiii passes throu«>1i a |K>orly il.'hned inner luwhin.l bordorod on the wes^t by a brerteheil and irreguhi'r (tuesta. Kastward, howniT. tht- dissection .>!' the c-oastnl phiin is niarkejl. and troHod hy the nnderlyiii"- neoUmv strnernre. on beils nt nnequal liardness. Idle most prominent feature dovehifwd m fhn sernieircvihir euosta tliat bt'^yins on the sontheost at l>e«tfni ffill mul extends northwest, then, west, passing^ south of Saijit John estate, thr-n.-e ^ontlnvestward to rh,»nnc Ksperani-e. 1'he escarpment rises 150 (pet above the low phiin on th»» northeast to an avera.fi-e ehn^ation of 500 feet and, att;iins n nnuirnnni heii^-ht or 5!)0 feet iinintaliatoir nortli ol' Buhwminde. Seen in profiJe. the encsUi appears as a series of ben«:hes or steps (Fii;. to), the most proinnient of wdiich lio af.proximatelv 350 feet above soadeveh The MF.YKRHOFF. I'll ysKHlRA PlIY OF YIIHHX isf.AMhs }(;; bt'ijcheh; iirr- rnp|M'<l bv the more rcsistam si'nurrv>talli!M'- litticsUmos ami jHrrmm but welJ-t-cnsolidatwl calt-aivDyto, wliib> t\w iiiarl^^ nmi oarthy liirie.stuiips J'uriii the gmth' iiiterniodiati' #Iu|h'>. Oit tin- buck ut' tbis semi- eirciilar <'U('sta tbo draiiuigi,' i^ radiall}- inwariU but iiotwitbstamiiii.e- tbc coiiccMitration of tut^oI? toward the <-c'iiti'r of tbf basmal stna-lurc. m> doriiiiiant cfiitral valley bas h\m lonm-.l. In^tejul, .-arh .-oiall valli>y inaintaiiiH its irii'U>pt'iii1fiicr and opens out on a rolIiti.H plain witb a biw. varyiii^f iiudination toward the soutb. The fUily strong- vaUey b'ns sontb of tbe eut'sta. which it trnncate.s. It trend.< sDiif hwc^lward tbron.«i-h snlxtnc'd cuesta-like vailey-walls between Anna's Hope and Work and liesr estates, appearin<{ to have been eroib^d npiai a bnv. poorly defiin'd anticlinal rf)ll in the Tertiary i^traia. It has assumed jiartiaj c<intro| of rbe cense(pu'nf (h-aina«re on the back slope of tbe l)a>in north of it. Tbe curve,! l.nvbiiHl at Idle ijase of the mcsia (Tart 1. Fi"-. 2} is not an inn<'r hnvland, for its surface is underlain by Tertiarv strata whose thickness and weakness fa(dli(at;ed tbcir deep and c^tensiv.' exeavatnuu iietweeri Lebanon iind Mornincc:b,ir estates in fact, as far I'ast a,s Saint John estate a broken hue of low Tertiary hills rise'< on the northern rim uf the hmiand. Tbey represent a dissccrd cnesta, overlook in- on tbe south the broad phiin at tb«> base r.f the nniin ciu'^ta. on the norib a d)nrJnyti\'e inner lowland whieli is little ni..re than a series of short ravines erofld m the ba,se of flie rreiae.i.ns upland ( Fi.l'. HU. The northern em-^ta ha> liecn In'eached Ice even tbc'leasf itnnortanr of the ;|jjH Si'lESTIFIV sVRVF.Y (IF THF VIIHHS IshASIlH iiitcriiiittciit nHntntHiii strcsitiis, which cut throii^uii it: mu} onter the Oh; (h*i«£iiHti<>ii. cojjsif^ts nicn-Jy of a sn<'r«».<Kioi) of di'v mviiies trihutrtrv In the hirgt-r valh-ys .lfVph»|H'cl jH.rrual to its tn-n«h .\'urth<Hisfwrtrd it afijuircs Uwnl iiii|mrtiiiH-(% whore Srtlt IJivi-r Hwidentallv tapped it as it .Toded hcndward toward tlic hir.s--,'r lowhoid en the suiitli *( Part I, Fi.sr. 'l). FM.twi>«.|i the «'m'>tHd,uwhind swtiojj of the coastal phiin and the ohi (daiii in the west, ihcn- is i\ »-ciitrally hn-atctl portioji of moderate relief. Its dissection has been act-oniplishwl hy the F.ethh;.hciii-Fair Phiiii stream system, th»' hirgcst drainage unit on Saint Croix. Diirin.^' tlie fielil, survey for the present report all the jueinherH <d' this system wiu'c dry except the up|«'r portion of tlic main valley ^, i^ whi<;li there was a small. irrejiijlar flow a, short distance above and belo\r K'iver estate. Mt^st of' the drainage J'nnn the easlern halt' of the .ddland north, ^ri the coastal plain enters this stn>am systtMii : <'ojnparativcly litth- oJ' it enters the Salt Ji'ivcr system, which drains the Tertiary lowhrnd described above. In the central coastal-plain area, in faid, the identity of the lowland ha,s been h)st by the removal of most of the higher 'rorliarv beds, and the st.n'ains nnike their wax a<-ross the sea ward -si oping plain in shallow conscqncnt valleys. Tlu- licthleliem-Fair llain valley apparently owes' its location to erosion on tin* crest of a very low north-sonth arch. t)u the .'ast .it is llaid«'d by Kingshill ridge: on the west, hy a ridge tliat is hT.iken sonth of (ndden (J rove estati^' by a tributary valley. Tlie tributary has captnreil the llivor-Castle Bnrke draifutgc whiidi formi'rly MEYERHOFF, PHYSIOGRAPHY OF YIRdrX ISLANDS 169 flowed due south through the wind gap near Fox estate, and has uiore recently extcuided its conquests farther west to the intermittent stream which flows out of the lowland past Hard Labor estate. The latter originally passed between the Tertiary hills upon whicli Adventure and Mt. Pleasant estates are located. Further piracy appears imminent ; for, in the soft limestones in wdiich the ill-defined valleys of these inter- mittent streams are cut, stream capture takes place easily. The Bethle- hem-Fair Plain system alone is firmly established and Avill naturally be the beneficiary. On the northeast Salt Eiver is a rival whose future development will initiate a struggle for supremacy between the two ex- panding systems. The divide between them at the present time is low and vaguely defined, consisting of a plain more than half a mile square in which there are not twenty-five feet of relief. In the preceding discussion the coastal plain has been considered mider three natural divisions: an old plain on the west, a mature to old plain centrally located, and a maturely dissected area of moderate relief on the east. The divisions merge: there are no clear boundaries. Ex- planation of the different degrees of dissection wliich characterize these divisions is not far to seek. From the Cretacepus upland north of the coastal plain, most of the run-off drains south upon the coastal plain. It follows that the amount of run-off* reaching the coastal plain is roughly proportional to the area of the upland ; hence where the upland is broadest (on the west), there has been a maximum amount of run-oft* upon the coastal plain, and a maximum amount of dissection has thus taken place. With decrea.se in the widtb of the upland eastward, the amount of coastal-plain dissection is less. Where the northern upland ends at Salt River, all of the drainage is indigenous to the coastal plain and comparatively little erosion has taken place. As the amount of decrease in the width of the upland is regular, so, too, is the proportion of decrease in the degree of dissection. While the coastal plain of the w^estern half of Saint (Voix w^as under- going fluvial denudation, the corresponding deposits on the north and south sides of the eastern half of the island were completely stripped from that part of the lower erosional surface which is at ])resent emergent. Erosion w^as not restricted to the coastal plain, and it may be assumed that valleys in the oldland were re-excavated, deepened, broadened and extended at the same time. In the west the rate of valley-deepening in the hard Cretaceous volcanics seems to govern the rate of removal of the coastal plain. Most of the valleys are accordant on both sides of the contact between coastal plain and oldland, but rapid removal of the Tertiary north of Predensfeld has caused aggradation. Gravels extend 170 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS from the mouth of Canaan valley out upon the coastal plain to Mon Bijou and Glynn estates, and alluvial deposits have been similarly formed elsewhere. The Submarine Platform The submarine platform surrounding Saint Croix is the drowned border of the subaerially dissected coastal plain. It is less than ten fathoms deep except beyond the eastern end of the island, where it ranges between 10 and 20 fathoms, with rea4ings more than 20 fathoms at its eastern tip. Along the north coast from Ham Bay to Salt Eiver Point and southeast to Long Eeef its greatest width is 1000 yards, with an average of 400 to 500 yards, while off Maroon Eidge it is prac- tically absent. On the west coast, from Ham Bay to Southwest Point, it is barely 600 yards wide. Elsewhere around the borders of the island its width varies from one to three miles, and beyond East Point it extends east-northeastward for nine miles, narrowing gradually from five and one-half miles down to three and terminating in a bluntly rounded end (Plate III). On all sides it stops abruptly at an escarpment 1000 to 6000 feet high, which sharply delimits its seaward margin. The character of the platform surface does not differ greatly from place to place. It has a very gentle, and in most localities uniform, seaward slope, which is broken by many reefs and shoals, and in the east by a few large banks. Two oldland monadnocks. Green Cay and Buck Island, rise above it to the surface, the latter attaining an eleva- tion of 330 feet above sea-level. The reefs, both growing and dead, include barrier and fringing types. Growing reefs extend almost con- tinuously from Southwest Shoal, about one mile off Long Point on the platform south of the island, to East Point. Beginning as barrier reefs a mile offshore, they gradually close in upon the shoreline, which most of them fringe from Watch Ho Point eastward. Small fringing reefs are present even along the northwest coast, but from Salt Eiver Point eastward coral growth is almost continuous, being especially abundant in the shallow water surrounding Buck Island. On the south- ern platform an irregular line of dead reefs rises 15 or 20 feet above the bottom. Off the west coast the narrow platform is unmodified. The banks which rise above the surface of the platform east of Saint Croix differ from the reefs in having larger dimensions and in rising from greater depths. They are found near the margin of the platform almost completely enclosing a submerged basin six miles long and two to three miles wide. Lang Bank, the largest, begins two and one- half miles north of East Point, extends for six miles along the northern MEYERHOFF, PHYSIOGRAPHY OF VJRaiy ISLANDS 171 rim of the platform, there bending south and ending at the southern rim. Its greatest width is less than a mile. Near the southern margin of the platform three corresponding banks, without names, extend from the latitude of East Point toward the southern end of Lang Bank. Two of them, each two and a half miles long by lialf a mile wide, are separated by a low gap scarcely 25 feet lower than their crests, which lie 6 to 9 fathoms beneath the surface. The outer slopes of all the banks are moderately steep escarpments to a depth of 20 fathoms or more, tlien very steep to 200 fathoms or deeper. East of Lang Bank the platform continues in a crescent-shaped terrace one mile broad and 14 to 29 fathoms deep, rimmed at its outer edge by narrow breached ridges (Fig. 47). Within the banks, on the other hand, there is locally a low escarpment at the crest, but generally the surface slojjcs evenly and gently to depths of 16 to 19 fathoms in the center of the basin. The submarine platform off the western, southern and northern coasts of the island offers strong contrasts to that of the northern Virgin Islands, in that (1) it is much narrower; (2) it is shallower; (3) the beaches deepen gradually offshore,—there are no steep slopes or escarp- ments (the northwestern fault-coast is excepted from this statement) ; (4) the modifications of its surface are restricted to reefs, broad ridges or banks being absent. East of the island, however, the platform is not essentially different, except in shape, from parts of the northern Virgin Islands platform,—for instance, south of Saint John. Points of resem- blance are found in the width and the depth, and in the flat-topped banks rising above the surface of the platform. The analogies afforded by this eastern area are sufficient to justify the conclusion that the history of the entire platform was closely parallel to that of the more extensive platform surrounding the northern islands. The nature and origin of the contrasts found in the other parts of the platform will be .con- sidered first. The platform off the southern coast, between Southwest and Watch Ho points, is an extension of the old plain of erosion developed on the coastal plain on the adjacent land. Its angle of slope is less, but the decrease has resulted in part from slow accumulation of coral sand and of debris washed from the island. There is no escarpment between land surface and platform except the low cliff cut recently by waves. Continuity of the coastal plain beneath the platform is the chief cause of the form which the latter exhibits. Deposition upon its surface, although con- siderable, has not succeeded in greatly modifying its character. Bast of Watch Ho Point the depth of the platform is four to six fathoms immediately outside the line of reefs and increases offshore to depths 172 SCIENTIFIC PURVEY OF THE VIRGIN ISLANDS greater than those found westward. The tendency of the platform to increase in depth toward the east is significantly related to the offset of the oldland to the south and to the contraction in the width of the island. The Tertiary coastal plain deposited upon this portion of Saint Croix was, after uplift, more deeply dissected than that to the west, and the greater depth of the platform is probably the expression of the greater dissection and perhaps of slightly greater submergence. Similar generalizations might be made for the platform off the north coast from Long Eeef to the latitude of East Point. Its surface has been considerably built up by reefs and accumulations of coral sand, especially in the vicinity of Christiansted and Buck Island; but else- where its surface, six to ten fathoms deep, is smooth and 'gently undulating. These portions of the submarine platform constitute a j^mng marine plain of deposition, apparently built up upon an old plain of fluvial denudation. Lack of any features of marine erosion elimi- nates the possibility that wave-cutting was responsible for its formation. Whatever work has been accomplished by the sea is depositional in character and has occurred comparatively recently. The eastern end of the submarine platform (Fig. 47), although it is modified by accumulations of coral sand, preserves all the characteristics of a fluvially dissected coastal plain. The horseshoershaped rim of banks, with their steep outer slopes and gentle back slopes, invite a com- parison with the Beeston Hill-Bonne Esperance cuesta in central Saint Croix. The outer slope is more than a debris slope, because it is strongest at the eastern end of Lang Bank, where it rises above the comparatively flat crescent-shaped terrace described above. It is not a wave-cut cliff, because in the same locality an outer line of ridges would have prevented wave-cutting at its base. In every particular except the ac(*idental character of its shape, it resembles the submerged coastal-plain ridges south of Norman and Peter islands on the northern bank. The basin within it attains a maximum depth of 19 fathoms near its center and 16 fathoms near its western end. Recent deposits have filled what- ever outlet the basin may have had, for the deepest channel leading from it is but 13 fathoms deep. It is doubtful if the comparison between the banks and the Beeston Hill-Bonne Esperance cuesta can be pressed very far. The latter owes its origin to arching along two axes tangential to the escarpment. There are no grounds for assuming that the entire eastern margin of the bank was upturned. Furthermore, if the platform has been dissected by fluvial erosion to the base of the banks (25 to 30 fathoms), denudation would normally be greatest at the sea border, as is the case of the cpastal 'f^ntoiireil ."iirfme nf the suhmmint philfurm mat of Stlril eroix, fthmi-mn mwjimd hatikt (mil MEYERHOFF, PHYSIOGRAPHY OF VIRGIN ISLANDS 173 plain farther west. Under ordinary circumstances the banks would have been removed unless the material composing them was of exceptional •durability in contrast to adjacent materials to which the run-off could have been diverted. Anegada will, perhaps, afford an illustration. It is forming at the present time near the northeastern margin of the northern bank, in a position and under circumstances analogous to those obtaining at the eastern end of the Saint Croix platform. Anegada and Horseshoe Eeef are growing upon a drowned unreduced ridge of Terti- ary coastal plain. There seems to be something inherent in the north- eastern ends of the submarine platforms favoring the growth of coral reefs and the preservation of the ridges which they build. The favorable factor for reef development is the location with respect to the trade winds, currents and waves. The most vigorous reef growth character- istically takes place to windward in the trade-wind belt. The favorable factor for coral-ridge preservation is found in the make-up of the reefs, which, when formed into rock, consist of a porous aggregate of coral heads, fragments and sand, firmly cemented with calcium carbonate. Porosity and firm cementation are qualities which give a rock excellent resistance to weathering, as the use of Bedford, Indiana, limestone for building amply proves. The eastern margin of the Saint Croix plat- form has thus been a favorable location for the development of resistant ridges. The fine debris derived from the ridge by wave clastation has settled within the ridge, as is illustrated again in Anegada. Away from the source of calcium carbonate cement, the fine clastic material is more or less dense, impervious and poorly cemented—an easy prey to fluvial erosion. There is, therefore, reason for believing that the eastern end of the Saint Croix platform satisfies every condition necessary for the development of hard marginal coral ridges with soft materials within. With an uplift of 30 fathoms, fluvial erosion would excavate the non- resistant deposits inside the reefs, but would reduce the reef-ridges very slowly. In this manner, the writer believes, the features of the platform east of Saint Croix were initially developed. Question might arise whether the marginal banks and interior basin would not have the same form if unmodified by fluvial dissection, and whether they could not represent original features of the Tertiary coastal plain as they are. The first question must be answered in th^ affirma- tive. The development of Anegada in recent time has produced a simi- lar group of features on the northern Virgin Islands platform. The second question must be answered in the negative, for deposition of a Tertiary coastal plain in this area must have been as rapid as it was westward, if not more rapid, because of the conditions which encouraged 174 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS marine life. At present the ridges lie 25 or more feet beneatli sea-level and are probably not more than 150 feet thick, compared with 550 to 600 feet of coastal-plain deposits at Bulowminde. Rednction by subaerial erosion is the only explanation; greater submergence is a contributory bnt inadequate factor. Unlike Anegada, the ridges do not now support actively growing reefs ;^^ yet, since the submergence which followed their dissection, a considerable amount of material has been deposited on the floor of the basin within them. This material could not have been derived from tlie adjacent island, yet it could scarcely have been derived from the ridges by marine erosion at tlie present sea-level. Again, an analogy must be sought in the northern platform. At the southern margin of the latter is a drowned reef whose growth was connected with a slightly lower sea-level during late Tertiary time, and whose death Avas caused by exposure during the Pleistocene. The submerged ridges on the eastern section of the Saint Croix platform rise to a fairly uniform height of six to nine fathoms, as if they had all formed at a wave-base five or six fathoms beneath the present surface, either by coral growth to that lev^l or wave-planation to it, or both. Either process would have supplied a source of material to account for the deposits in the basin. There are other grounds for postulating a temporary wave-base at a level some five fathoms beneath the present water-line:— (1) North and south of the eastern half of Saint Croix the platform maintains a level of six or more fathoms immediately offshore. Unlike the platform surface to the west, this area apjiears to have been formed at a lower level and has not yet been built up toward the present w^ater surface by accumulation of clastic material. (2) Of! the northw^estern fault coast, where a narrow plat- form is present, it is regularly five to eight fathoms deep. It appears to be a wave-cut terrace whose formation was possible only when the waves were cutting at a lower level on the island's flanks. (3) Finally, the features of drowning exhibited in the coastline of the island are. such as would form with a small amount of submergence and not with the total amount of submergence represented by the present depth of the eastern part of the platform. The final problem w^hich the platform presents for solution is the origin of the differences in its depth and character at the eastern end and in the portions flanking the island. The gradually increasing depths of its surface "from west to east, beginning approximately at Watch Ho and Christiansted, suggest downward tilting of the eastern s'' Notations on the cbarts indicate that there is some scattered coral growth on Lansr Bank, but It has not jet assumed Importance. MF.YERHOFF, PHYSIOGRAPHY OF riHOlS JSLAXDS 175 portion. The decreasing elevation of tlie Cretaceous npland in the western oldland area apparently supports this view, although the more thorough dissection which it has undergone is an equally plausible expla- nation for its lower altitude. Tlie nature of tlie coastal-plain dissection south of Christiansted indicates tliat arcliing occurred in this region ; it would follow that depression occurred to the east, but whether it was enough to explain the features of the eastern platform is not known. Furthermore, the deformation seems to have preceded dissection of the coastal plain and in that (^ase would not bear on the prol)lem. The only alternative hypothesis which can be offered to explain the difference in the eastern and western platfonn is slight down-faulting; but its case is weak. There are not suffici(»nt data for a final conclusion. From the evidence considered, the history of the submarine platform appears to have been a phase of subaerial coastal-plain dissection, fol- lowed and complicated by submergence and marine deposition. Be- ginning with the formation of the (Mjastal ])lain upon tlie lower erosional surface of the oldlaiid, the balance of its liistory may be summarized as follows:— (1) uplift and fluvial dissection of the coastal plain, with the production of an old plain on whose^ eastern surface a few residuals were left only partly reduced ; (2) partial submergence, followed by reef- growth and deposition; (3) temporary emergence during the Pleisto- cene; (4) resubmergence to the present level. The Pi?esext Level The principal features which liave developed during the sea's present stand include continued dissection and removal of tlie coastal plain, the development of the shoreline, and the growth of coral reefs upon the submarine platform. Dissection of tlie coastal plain has already been discussed ; attention may, therefore, be restricted to the other two fea- tures. In its relative simplicity the shoreline of Saint Croix offers a strong contrast to the irregularly distributed, deeply embayed and oddly shaped islands of the northern Virgin group ; yet it is the result of a series of events probably no fewer in number than those responsible for the present forms of the northern islands. The reasons for its apparent simplicity are many, but those of chief importance are closely related to the different varieties of shoreline exhibited ; hence the types of shore- lines and their synthesis to produce the composite outline of the island will be considered jointly. The present form of Saint Croix represents the synthesis of three types of shorelines whose origins are dissimilar. The only features they share in common result from recent slight submergence. The scmthern coast 176 HVIEh'TlFlfJ i<rRVEY OF THE VJRGIN ISLANDS from Sandy and Soiitliwest |)oiiits eastward to Watch^ Ho has been formed by margijial svibmergence of the fluvially dissected coastal plain. By the processes of (k'nuchition the higher strata of the coastal plain were removed witli considerable uniformity over much of the surface adjacent to the coast, especially west of Krause Lagoon. The surface has little relief, and even the intermittent streams have eroded very shallow valleys. Their courses from the upland are less incised south- ward, and for tlie most part tlie southerly slope of the strata near the coast is so uniform that only locally does the run-off show a marked tendency to concentrate in wcll-defiued valleys* The result is that the run-off is commonly little more than a sheet-wash during heavy raius, and erosion goes on at a comparati^'ely uniform rate over the entire seaward portion of the coastal |)lain. Because of the absence of strong valleys, drowning of the coast lias not produced the usual deeply em- bayed shoreline, and only a" few localities offer positive evidence of sub- nu^rgeiice. Drowning has produced a diminutive bay, into which drain the two intermittent streams entering Long Point Bay , about half a mile north of Long Point. The Bethlehem-Fair Plain stream widens for over half a mile above its mouth. When the locality was visited in April, 1924, the stream bed was dry, but the lower valley was filled by more or less isolated pools of salt water for more than a quarter of a mile from the shore. l)cj)ression has been responsible also for cutting off the meander beliind the l)eacli. The width and arc of the meander are not in 'accordance with the greater width of the present estuary— a feature to be expected if the mouth of the stream has been widened by the influx of salt water concomitant with drowning. Wherever the sea has found higher strata to work upon, it has cut low cliffs, and the entire southern shoreline of the coastal j)lain is a series of sand beaches along the lower portions of the ])lain, interru])ted by low wave-cut cliffs whose height is rarely more than ten feet atid never as much as twenty feet. The best are found in the vicinity of Long Point and at Cane (Jarden Bay, where Tertiary hills of moderate elevation front the coast. The irregularities of the shoreline, althimgh of small dimensions, have under- gone so little modification that tliey still possess all the characteristics of youth. Tlie form of the scnithern shoreline is considerably modified by fring- ing and barrier-reef growth. Krause Lagoon is partly enclosed by a barrier reef, the eastern half of which is exposed, the western half still submerged and growing. Sand has accumulated along the seaward margin of the dead portion ami is building up a cuspate bar outward toward a growing coral patch -lOO to 400 yards offshore. The position MEYERHOFF, PHYSIOGRAPHY OF VIROJN JSLAXD8 177 of the eastern half of the reef above water suggests recent slight emer- gence as the last stage in the island's history. Further evidence bearing on slight emergence will be discussed in following paragraphs. The Sandy-Southwest points peninsula requires special consideration. Tt is an anomalous protrusion from the southwest corner of the coastal plain, enclosing two lagoons, one of considerable size. The form of the peninsula is simpler than that of the adjacent southern coast and its origin seems to have been more recent. It rises less than twenty feet above the level of the sea and is composed of firmly consolidated lime- sand, bordered seaward by a beach of unconsolidated sand. The form and nature of the peninsula indicate that it developed by continued out- growth of bars into an advancing double cuspate foreland, which has since been upraised. The material composing it is unfossiliferous, and its age cannot, therefore, be stated ; but it is thought to be Pleistocene. The consolidation of the lime-sand is a normal process occurring at the upper reaches of the waves along all the present beaches, where by evapo- ration of the water in the pore spaces between grains a calcium carbonate cement is precipitated. The result is a very solid rock, firmer than most of the poorly unconsolidated Tertiary marls. It seems pro])able that the peninsula was formed by the forward building of bars through the agency of confluent coastal currents. The interbar depression was ulti- mately cut off from the sea, and West End Salt Pond was formed. The elevation of the peninsula above sea-level is the result of emergence, possibly connected with the formation of the structural arch whose axis," according to Quin,-'* extends from Saint George Hill to Southwest Point. The inner shores of the lagoons approximate the form and position of the older coastal-plain shoreline before modification (Fig. 48). The western shoreline is exceedingly regular in outline. At the south it extends from Sandy Point in a broad sweeping curve until it joins the older Tertiary shoreline south of Frederiksted. A beach of coral sand borders the consolidated material of the peninsula, whose low, flat surface is not more than fifteen feet in elevation. At the south end of Frederiksted there is a low sea-cliff less than ten feet higli, but in general a cliff is absent where the western shoreline is formed upon the coastal plain. The coast curves broadly in the vicinity of Frederiksted, and thence trends X. 17° W. in an almost straight line two miles long. At Sprat Hole the Tertiary ends against the Cretaceous voleanics, which border the sea northward from that point. The character of the coast changes abruptly with the change in rocks, from low and flat to rugged and high; for the oldland rises rapidly to elevations of 750 feet within 'Quin, J. T.. Op. cit., pp. 44-52. i:h Si'lHSTIFh' HVUVEY OF Till-: YUHilX ISLAMhs a mile t'rom tlie >^horc'. Sliiillow re-ejit rants aii«l Irnr wrtve-mit elitfs bt'i'uiiif soiriewhiit moj'c iiiinioroiiH, but for tlio won part there is a bcaeh of iiiH-oiiH.li.bit(Hl iilliiviiil JiuitpriHl. whosP <lfvul|mH'nl luw eoiK^iderMbly iiKidiOfd the r,rj.i«iiial ii;r«',i5-ulariri«'s. 'rin> extent of tin- ninililifsiiioii ciiii be a|)|>r<'ci:it«>il bv wmtrnstio.i: t!ie prrsinit sh(n'(>line wjtb, t\w linndrcd- fdi.it (-(intimr ( Fi^^ 10). Tlicre is nsiiallv a steep but rc^aiilar rise from tho s«>a to HJi t'lt'vatioti of twcnity JVet, above wbieh the slopv <£rmluii]ly assumes tbe form eharai-teristic of fluvial <lisseetioii. With the jiassage MErf]iiijiH<'F. J'lrysioiiRAf'HV df ruiaiN iSL.ixihs i/^i oC the .<hf»ivlii!o rruin the coartal plain to the old- huid. there is a eliang^e in the width of the snhiruv riiie platform otfsh,<>i'e rr<i!ii an average of (UM ynnh to iiiM yanb. The widtli of the phitforiu above tlie len4:ath(nn edi- toiir see ins tu he n fair measure of the aiiioimt of marine ])lnnation, whieh has taken phe-e siuee the western nuirgin of tlie platform was formed : thns it is witler on the easily eroded materiid of the cortj^tHl plain, bnt nar- row where it ail joins t!ie old hind, whieh, was re- dneed ini>re slowly he- eatiHe of its hei<j[ht and tlie Inirdriess of its roeks. Coral growth has not ob- tained a foothold on this sitle of the island, an<I aeemnnlatitm of material Ijas been a nepliicihle fat- tier in nn.dilying the snh- nierged surface. The north <-oast from tlani PdulT to Baron Pdnff (Fart J. Fill'. :!), a dis- tanee nf six nriles. ollVrs an interesting^ mixture of forms. Along a nnle of it. between Ham PdiitT antl Annaly liay, no snhjme- rine terraee is prj'sent.hut ^^.^^^ ^^ the stei.|. roek walls of Mnrnon Ki.lge (Fig. V.l) T*'^^^^ |s,i) sriHXTiFif; Hrh'vi-:Y of thi-: \ifH;ix /.vljjd.v* <l(%s<t'inl ijito the ofi'iiii, with liHU' rhaii.yx- or nngU\ to depths <.f lOi) ratii^ mm. 'Thr ^^horeliJic U in cxtrciiH' vuiitli rtiid has Ih-cii Jillh' more thiiii iH>t<-h(N|: it is <:!)nrM(;tcri7,i..l hy hi«:'h Kca-clifl's which '.m-. cumph'tely iri^ iU'ci'Hsibl.' fVom hind ur sea, wlrih« iibuvc tho watcr-liiie tht- IhiiikH (d^' the rido-e art' ^^(•aI<•(dy ^nilli^'d. I'^asiward fnun Aniiiily Bay the coast changes. The ti|dand is sainewduii dissected : ami short, !*teep vaJhw^ (h^'scei)d to th«^ sen ( Fi^'. ">'')? torndrig at th(4r mouths sha]h)w rc-entrants, ahaig which are iiarmw heatdies I Part I, M"!^. 3). iHstweeji the small eni'havnicnt- Miv hi«,di rocky stretches, usual ly cut into ini]MssHl)lc sea^dilTs. In places the pres^'Hrc of a narrow roek bench a little above mean tide h>vd, hut within reacb of the; higdier waves, .^u^-gests ree-ent sli^fht npiil't of ttic coast; but the extrcnte difficulty of travel alon.a' ninch of tlie short- line made careful i^xaTtdnatioii of the nature atnl origin of these benches impossible, and the ipiostltai of their sigid(icance. however snggestive oh cim'rg-eiici., cannot be regarded as settle<l. The suhjiuirine platf<»rm along this portion of the ee»ast is .narrow but constant, varying in wddth between lUni and ')iM» yards. From Wells Bay eastward l»ca(dies hecmnc^ nnu'c nntnerous, while wave-cut clitTs are reslricted to the nnire pronu- nent. hcatllands, of whicdi lianm Bluff is the easternmost. Between Lavallee and Ihistoplwisf estates tdiere is a hr»ad flat, filled witli allnvial deposits, which have hw'ii waslied down from the upland in a network- id' short, steep, converging valleys and spread out between Ihe ridges on either side. 11ie nuiferial has nearly filled what was on«-c tlic largest enihayuH'ut along this ,,5,rt (.f ibe north coast. The nn.de of develop- nn>nt was Ihe nornufi one: a bar formed across tlic nnmth of a re^/ntrant. ;illifviiini filled flic landward si<le. an,d in between there was fornuMl a IIEYERHOFF, PHYi^JOaKAmV OF \ IRdLX /NL 1 V/>.s^ 181 lagoon, which ]iow holds water only in rainy niontlis. Alluvial filling is, apparently, almost completed. Anotlier arresting feature of the north (/oast is an asymmetrical, triangular facet along the coast west of Wells Bay (Part I, Fig. 8). A spur from the ujjlaiid is truncated hy the (^oastline and descends to the sea at a 40° angle from aii almost knife- like crest 895 feet liigh. It has heen postulated that the northern coast from Ham Blutf to Baron Bluif is a young fanlt-scarp. The hypothesis is fully home out hy the physiographic evidence. Tjie following are the critical points: (1) the limited development of the submarine |)latform, ])resent as a narrow strif) where the dissection of the upland has ])ermitted the short, youthful valleys (Fig. bO) to carry and de])osit land waste along the shore, absent wdiere the upland has undergone least erosion ; (2) a shore- line in its initial to youthful stage of de\'elo]unent (Part 1, Fig. 3) ; (3) the narrow watershed on tlie north side of western Saint (^roix, as contrasted with the broad w^atershed on the south side of the dissected oldland ; and (4), in consequence of the asymmetrical divide, sliort, steep valleys on the north (even these are a])sent along Maroon Eidge, Fig. 43), in contrast to the longer, more mature valleys on the south. Oue (4iaracteristic ])hysiographic feature is missing: a rectilinear slu)reline. Few l)l{)ck faults are simple, however. Th(» fault boundary of the ITigli- lands in southeastern New York, for example, consists in places of a series Of short faults with different strikes occurring in a zone which has a linear trend. The faulting north of Saint (*roix may, therefore, be zonal and somewhat irregular, the outline of the (*oast conforming with the irregularities. The northern shoreline between Baron Bluff and Christiansted is jiybrid. Salt Eiver has bo^en drowned and dismembered, and, unlike the other drowned river mouths in the Virgin Islands, the estuary is a com- paratively broad, open one (Part I, Fig. 2), two fathoms deep, and only ])artly silted up along the sides, where short intermittent tribu- taries enter it. Filling seems to have been in })art retarded by the screen of mangroves which grow arouiul tlie edges, and in part pre- vejited by the perennial flow of Salt Biver, which maintains an open channel into the sea. Reefs are growing across its mouth, and sand derived from them fills the outlet to a depth of one fathom, but com- plete closure (^annot occur. I'he shoreline southeastward from Salt Iiiver Point to Christiansted consists of unconsolidated sands nnirgining the rocks of the oldland at the north, a narrow strip of the coastal plain in the central portitm, the oldland rocks again in the vicinity of (4iris- tiansted. Tn the neighborhood of Judiths Fancv the otherwise low 182 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS coastline is interrupted by hills of folded Cretaceous volcanics, which front the sea in wave-cut clitTs. I'he isolation of the hills, which are cut off from tliose westward l)y a fiat plain of poorly consolidated, partly reworked material of Tertiary age, indicates that they were Tertiary islands, formerly co\ered by the coastal plain, which has since been stripped from them. The rest of the shoreline, from Christiansted on the north and Long- ford estate on the soutli, is a physiographic unit. It has been formed by the submergence of the maturely dissected oldland with the formation of embayments which \ary between one and two and one-half fathoms in depth (Fig. 37). Altona Lagoon, the largest one maintaining an open connection with tlie sea, was caused by submergence of an east-west valley. Other major features are the broad basins in Avhich Southgate and (ireat ponds are located, initially formed by rapid erosion upon a granitoid intrusive, which probably underlies the greater part of the two areas. Submergence caused the deep embayments, bars developed be- tween the headlands bounding them, and alluvium has filled a large part of the depressions. The lagoons situated between the alluvium and tlie bars are large, but are only one to two feet in depth, and in dry weather Great Pond sometimes dries up completely. Both lagoons are, therefore, all but filled. Farther east the island narrows, the valleys are shorter, and initial re-entrants were, in consequence, not so large. The bar- lagoon stage of development was omitted in the shortest valleys; and in several instances where lagoons formed, they have subsequently been filled, only four small lagoons remaining unfilled in this part of the island. Sandy, evenly concave beaches alternate with long stretches of rocky, wave-cut promontories, which in many places tower more than fifty feet above the sea. As much as half of the coast is composed of wave-cut cliffs. Coral-reef growth upon the submarine platform has already been men- tioned, but a resume of its nature and significance is pertinent to the features of the present level. Along the northern fault-coast coral growth has barely obtained a foothold. Scattered heads have acquired a footing among the rocks south and east of Ham Bluff, but they cannot l>e classed as true reefs. A thin, fringing reef growth begins on the north coast at the mouth of Salt Eiver and continues in patches to Big Princess estate. Thence Long Reef, barely submerged, extends away from the shoreline at a low angle, continuing subparallel to the coast in Scotch Bank. The reef thus formed encloses Christiansted "harl)or,'' a small anchorage from two to five fathoms deep, between Protestant Cay and Fort Louisa Augusta, to which access is obtained through a natural. MFAERHOFF. PHY^lOORArHY OF VIRfUN h'^LANDfi 183 tortuous channel seven fathoms dee]). The bahince of the sand-filled lagoonal area behind the reef is less than two fathoms in depth. What- ever harbor Christiansted may be said to have owes its origin to the bar- rier reefs, which threaten to take that which they have given. The channel is so narrow and tortuous that only short vessels can make the turns, and for this reason the port of call for larger boats is Frederiksted, where anchorage is found on the ex|)osed bank. Coral growth is abun- dant but scattered between Green Cay and Buck Island, the latter being bordered on all sides except the southwest by a fringing reef, while irregular growth northward has formed a shoal more than a mile broad. The mainland to the south is also protected by a fairly regular reef, part barrier, part fringing, from Mary's Fancy to P]ast Point. On the south coast, a barrier-fringing reef extends with a few short breaks from East Point to Great Poml Bay, west of which coral growth is less regular. A drowned reef of older origin parallels the coast at a distance of two miles, extending to the western margin of the submarine platform. The reefs surrounding Saint Croix have furnished an abundant sup- ply of coral sand to the surface of the submarine platform and to the sandy beaches along the coast. The coral growth and derived clastic material have, in consequence, modified greatly the original surface of the submarine platform and have often been instrumental in producing terrace-like effects in places where sand and land-derived wastt^ have accumulated in horizontal deposits within the reefs. The deposits which are now being made are forming a new coastal plain, and their mode of formation is like tliat which produced the Tertiary coastal plain. In concluding the consideration of the present level a few summary statements appear necessary. The apparent simplicity of the shoreline of submergence is due entirely to the land forms which produced it. Part of the northern coast is a fault-scarj), which suffered little fluvial erosion before it was submerged to the present level. The southern coast is a partly stripped coastal plain, upon which subaerial erosion proceeded so evenly that no valleys developed to cause embayment. The remain- ing eastern part of the island is a re-excavated lowland margin, which was formed initially by early Tertiary fluvial erosion and was later covered by a coastal plain which has been removed. Most of the short valleys dissecting the remnant of upland were too steep for the forma- tion of deep re-entrants upon submergence, and many of those formed have since been filled by alluvium, which has obliterated the lagoons formed behind bay-mouth and mid-bay bars. Hence even in this portion of the island the shoreline has been greatly simplified. Modification of tlie shoreline by wave erosion is gradually coming to an end, because 184 SCItJXTJFlC SIK\ EY OF THE \ JRaiK l^LAXD^S much of tlie coast is now protected by fringing and barrier reefs. Their protective influence indicates that subaerial erosion is the dominant factor in denudation in regions where coral growth flourishes, and their effect as a check on wave erosion explains why marine planation has ])layed so little part in the physiographic history of Saint Croix. At the present time the reefs are not merely protecting the mainland from wave-attack, but the living and drowned barrier reefs are serving as wall^ within which a new coastal plain is being built up. It is possible that the present shoreline of svd^mergence has been modified by very slight emergence. The rock benches on the north coast and the elevation of the Southwest-Sandy points peninsula may be evidences of recent sliglit emergence, but this possible event requires additional confirmation. (^HAPTER IV ITIE FHYSKXiHAPHY OF VIEQUES AXl) ( IJLEBKA The Oldland and Its Peneplanatiox A. K. Lo})eck visited tlie islands of Vieques and (kilebra in the course of his physiogra])hic survey of Porto Rico in 1916-1917, and he has sketched tlu} nuiin features of their development and present form in connection with the physiography of Porto Rico.^^ In this chapter their liistory will be discussed in the light of the physiography of the Virgin Islands, with which their topographic forms are, surficially at least, more obviously related. Their history begins with an oldland of Cre- taceous volcanics and sediments which were intruded, folded and elevated. Peneplanation followed; and, although the even crests of the islands plainly represent a peneplane, Lobeck's conclusions raise a ques- tion as to which erosion- cycle was responsible for the flat upland to winch both islands rise. A brief summary of the relations of the two islands will assist in solving the problems of correlation. Both islands rise above the same submarine platform upon which the northern Virgin Islands are situated. Culebra, the, more northerly and easterly, lies ten miles south of the northern margin of the platform and eight miles west of the nearest of the Virgin Islands. It is sepa- rated from A^ieques on the south by an equal distance, from Porto Rico on the west by seventeen miles. It is flanked by a number of small islands: on the east and northeast by Culebrita Island and Northeast (^ay; on the northwest by several rocks and diminutive cays, of whicli Fungy Bowl is the largest; on the west by Luispena Cay, The Sisters, 2« Loltock. A. K. : "The Physiography of Porto lUco;* pp. .'^14, 320, 324. 3.10. 373-374. MEYERHOFF, PHYSJOGRAPHY OF nRdlX It^LANDS 185 Cross and Cactus cays. Virgin Passage between Culebra and Saint Thomas exhibits depths that are characteristic of the deeper levels of the platform, shallowing from 25 or 30 fathoms on the north to 12 anvl 1? fathoms between Culebrita and Savana islands. It is thus a normal feature of the submarine platform. The depths surrounding Cidebra average seven or eight fathoms immediately offshore, and are reached by steep but not precipitous slopes from the shoreline. Greater depths are attained within short distances of the coast, for the bank dro])s from the seven- to ten-fathom bench surrounding the island to 25 fathoms within one and one-half miles northward ; to 16 fathoms within an equal distance southward. Aaeques Sound, between Culebra and Vieques, averages 14 to 16 fathoms in depth. Vieques, situated one to two miles frcmi the southern nuirgin of the bank, is second in size only to Saint Croix among the islands considered in the present study, measuring eighteen miles in length and four miles in Avidth. Integrally related to it are five small cays located within half a mile of the coast. It is completely bordered by a beach which slopes gradually to the depths of the surrounding idatform. Off the eastern half of the island the latter is 12 to 14 fathoms deep both on the nortb and south; eastward the depths are greater, while westward and nortli- westward soundings of 8 fathoms or less are prevalent to Porto Rico. The rocks of tlie oldland on both ishnuls are of the usual types. Culebra is composed entirely of surface volcanics and shallow iiitrusives, most of them massive, but with a few thin-bedded, stratified nunnbers well displayed on the east side of Flamingo Point. The structure of the rocks is gently undulating but nearly flat. On Vieques the rocks dis- play more variety. From Isabel Segunda eastward to Porto Salinas and Salina del Sur, where they are overlapped by Tertiary strata, the rocks consist of folded, indurated surface volcanics with a lesser amount of interbedded crystalline limestones. Surface volcanics continue west- ward in a narrow strip, half a mile wide, along the north coast to tlio base of Punta Arenas. Except for this strip and a narrow belt of over- lapping Tertiary coastal plain deposits in the south-central peninsulas of the island, the western half of Vieques is underlain by a granitoid intrusive, not essentially different from that found on Virgin Gorda. The topographic forms produced by the two types of oldland rocks are strikingly different. The eastern half of Vieques is rugged, with a fine- textured topography in which the dissection appears to have follow(Hl structural lines. Examination w^as not sufficiently thorough, however, to establish any detailed relationship between surface features and geo- logic structure; but the strikes of the stratified material are approxi- 180 (^CIEXTIFIC PURVEY OF THE VIRGIN ISLANDS mately northwest, and it was noted that many valleys and ridges showed similar trends. Hill alignment and crudely parallel valleys and ridges are best displayed south and southwest of the abandoned ^^centraF' at Le Brun stack. The northwest direction of the Ensenada Honda shore- line is also suggestive of structural control, but the inaccessibility of southeastern \'ieques made a visit to that portion of the shoreline im- practicable. Tlie western half of the island is coarse-textured ; the surface forms are smooth and rounded, lacking the angularity that is seen in the \'olcanics to the east. (Ailebra rises to a maximum elevation of 650 feet in Mount Hesaca, which is located in the north-central part of the island. Dolpliin Head is 551 feet high, i)ut none of the other elevations are much greater than 400 feet. In Luispena Cay the central hill rises to a height of 476 feet, but the other cays have maximum elevations less than 350 feet above sea-level. Tu Vieques Mount Pirata at the western end of the island is the only conspicuous monadnock, attaining a height of 981 feet. From Culebra or from the mainland of Porto Rico, the skyline of the western half of Vieques appears remarkably even, and in the less reliable close- up view obtained from any of the hills east of Mount Pirata the accord- ance of the hill-summit elevations is striking. The highest is 594 feet, but nearly all of them approximate 500 feet. The significance of this constant level is in no way nullified by the fact that barely ten per cent, of tlie entire surface rises above the 400-foot contour, because the hills are well distributed through the central and northern ])arts of the island and are also represented in the southern part west of the coastal plain. Viewed from Culebra, the low, broken skyline of the eastern half of Vieques stands in striking contrast to the smooth regu- larity of the western half. Heights of 300 feet are prominent, and most of the passes cutting through the island are less than half that elevation. Only Mount Jalobre, west of Salina del Sur, stands at a height comparable with the hills of the west.^^ In his inter|)retation of the significance of the persistent higher eleva- tions in Culebra, Lobeck frankly admits the absence of decisive criteria but assumes that "the tops of the hills roughly indicate the upper leveF^'^^ of Porto Pico. Because in Vieques the Tertiary coastal plain overlaps a lowland cut below the level of the even-crested hills, Lobeck expresses the opinion that there also the hill summits represent the upper ]>ene- plane. From the point of view of the physiographer whose study wa^ 8»Cf. rx>beck. ()|). cit.. Fig. 40. p. ,'!74 : V. S. Coast Pilot, Porto Rico and Viri?In Islands, Dept. Commerce, U. S. Coast and Geodetic Survey. Serial Xo. 1(»4, 1021, Fig. 1. " r^olwck, A. K., Op. cit., p. ai4. MEYERHOFF, PHY^^IOaRAPHY OF VlRdlX l.SLAXDS \S7 restricted to Porto Eieo, such a conclusion is inevitable. In Porto Eico the upper and lower peneplanes are such outstanding features tluit inter- mediate levels of small area appear insignificant by comparison and are likely to be overlooked. To tlie Mrgin Islands pliysiograplier, witli small residual areas to study, the impress of shorter erosion cycles assumes greater importance because they affect a mucli larger pro})()r- tion of his province. Eastern Porto Pico imposes an additional difti- culty in the correlation of its land-forms with those of Culebra and Meques. In tlie monadnock group of tlie Luquillo Mountains it can luirdly be said that the upper peneplane is represented at all, except possibly in a few flat-topped spurs^ 1400 to 1600 feet high, protruding beyond the mountain mass and locally flanking the lieadwaters of the streams wdiich are working headward from the present baselevel. Dis- section has destroyed nearly every trace of the upper peneplane. East of the mountain mass from Cape San Juan to Humacao (probably farther, but the writer's observations were limited to this section), tliere is a series of low hills rising between the ])layas and constituting tbe stream divides^ which attain elevations of 400 to 600 feet. The acc^ord- ance of these hills is particularly imposing when observed from north- western Vieques^ from Palominos Island, and from the lighthouse at Cape San Juan. They are given scant attention by Lobeck and possibly too mucli by the autlior.^- Tliey do not represent tlie upper peneplane, which, if it is present at all, has only fragmentary representation m tin* s|)urs on tlie mountain flanks at 1400-1600 feet. According to Lobeck's |)hysi()graphic map of Porto Pico tlie Lnquillo Mountains droj) otf sharply at their eastern border to the level of the lower peneplane, whose surface is broken by numerous hills and ridges rising above it. Lobeck has not considered it significant that these lower hills and ridges approxi- mate a common level, although he recognizes tliat tliey cannot represent the upper peneplane. Yet, if Mount Pirata is excepted, scarcely a single hill on Vieques or Culebra rises higher than the general level attained by the low hills of eastern Porto Pico. If, then, the Vieqnes- Culebra summits indicate the upper peneplane, it wcnild follow that tlie latter is depressed 1000 feet in the ten miles or l(»ss which se|)arate Vieques from the mainland of Porto Pico. Plainly some other explana- tion than the tentative suggestion made by Lobeck must b" sought for the accordant summits of the two islands. The following possibilities suggest themselves : 32 These hills will be discussed more fully in a subsequent report on "The (Jeolosjy of lTje Fajardo District," to be published as Part III of Volume II in the Scientific Survey of Porto Rico and the Virgin Islands series. 188 HCIENTIFIC I^URVEY OF THE VIRQIN ISLANDS (1) That the upper peneplane has been faulted or flexed down 1)8- tween Porto Rico and Vieques. (2) That the even hill summits of Culebra, Vieques and the Xaguabo- Fajardo area represent a peneplane contemporaneous with the upper peneplane of central and western Porto liico, but lower in elevation l)eeause it was formed by short streams working into the oldland close to the coast. According to this view the peneplane would be developed at higher levels with increasing distance from the sea. (3) That the Vieques-Culebra upland, together with the hills and ridges of the Xaguabo-Fajardo area, consists of second-generation mo- nadnocks rising above the lower peneplane. Accordance of summits would result merely because of the subequal rate of erosion affecting all areas more or less alike. (4) That the (hdebra, Vieques and Naguabo-Fajardo ridges represent an intermediate peneplane equivalent to the second-cycle level identified in the northern Virgin Islands from discordant tributaries, broken main- valley gradients, interior basins and coastal rock benches. If this theory is to be accepted, it must account for the fact that the erosion surface, meagerly developed in the adjacent Virgin Islands, has an extensive development in this area. As eacli one of these interpretations has some plausibility, they require independent consideration, especially as Lobeck inclines to the view that the hilly interiors of Culebra and Vieques are to be correlated with the u{)per peneplane of Porto Rico. ( 1 ) The upper peneplane has been faulted or flexed downward be- tween eastern Porto Rico and Vieques. West of the Luquillo Xational Forest the upper peneplane is approximately 1500 to 1800 feet in eleva- tion, ending southward in the ridge, 2000 to 2500 feet high, Avhich over- looks the Caguas lowdand. The only locality west of the Luquillo Mountains where faulting might have occurred is in the broad, deej) \alley of Rio Canovanas, whose headwaters have even cut a comparatively low pass in the high southern ridge. But remnants of the peneplane are found east and soutli of the Rio Canovanas lowland, as noted by Lobeck, and are probably present in the spurs east of the Luquillo Moun- tains. Extension of the peneplane across this deep valley eliminates all possibility that it is a fault valley, east of which the upland might be found at a lower elevation. East of Rio Canovanas there are no topographic features, such as lowdands and rectilinear truncation of structures or of surface forms, wliich would suggest large-scale faulting. The eastern coast of Porto Rico is an irregular shoreline of submergence lacking the rectilinear outlines of a fault coast, and submarine contours MEYERHOFF, PHY^SlOGlxW I'HY OF YJR(;JX JSLAXI),'^ 189 between Porto Eico and the eastern islands reveal the phitform as an area of submergence whose surface could not lia\e l)een faulted. The intrusive diorites in western Vieques appear to be continuous with the rocks of the Yabucoa-Humacao area in the adjacent part of Porto Pico. Hence between Porto Eico and A'ieques there is neither a topograplric nor a geological break. Faulting, therefore, is not likely and is certainly impossible on a scale large enough to account for the difference of 100(^ feet in the upper peneplane on the mainland and the erosion surface in the islands. * It is true that the Saint John peneplane of the Virgin Islands is at an a^'erage elevation of 1000 feet, as compared with 1500 to 2000 feet for the upper peneplane in eastern Porto Eico. If the two surfaces were formed contemporaneously, as the writer believes they were, warping or tilting of the peneplane must ha\'e occurred. If it could be ])ro\'ed to have taken place east of the Luquillo Mountains, the explanation of the \ ieques-Culebra upper erosion surface could be simply solved. Such a solution might also be extended to include the Xaguabo-Fajardo hills, which are certainly comparable in form, altitude and tinu^ of origin writh those of the islands. But there is no evidence of flexure on the eastern flanks of the Luquillo range. On the contrary, a very close study of the region with this problem in mind offered convincing proof that the u])per peneplane is not wholly absent, as stated by Lobeck, '''•'' but is repre- sented in several isolated, deeply dissected spurs aud in one extensive area due east of El Yunque. The erosion escarpment between these upper flats and tlie Naguabo-Fajardo hill levels renders it certain that the latter represent a later and lower surface of denudation. Because of their apparent genetic relationships, the lulls of east<;M-n Porto Eico and of Vieques and Culebra cannot be explaine<l as features of diverse origin; therefore, monoclinal flexure of the u])])er ])eneplane between Porto Eico and the eastern islands is not admissible. Instead, the tilting which has depressed the island section east of Porto Eico about 1000 feet was gentle and even. The uplift wd\icli occurred after de|)osition of the Tertiary coastal plain was 500 feet greater at Lares than at Corozal,''* and at least 250 feet lower between Canovanas and Eio Grande than at Corozal—a total of 800 feet in 65 miles. Tilting of the coastal plain downward toward the east was accomplislied hy a subequal amount of tilt- ing in the up|)er peneplane. If the angle renuiins constant still farther eastward, the surface of the upi)er ])eneplane would coincide with the Saint John upland in the \'irgin Islands. There is no basis for assum- 33 Lobeck, A. K. : Oi). cit.. p. 314. ^* Lohet'k, A. K. : Op. clt., p. 327. 190 SCIENTIFIC SURVEY OF THE VJIWIN ISLANDS ing irregularities in tilting, and no reason, therefore, for including the upper levels of the islands between Porto Eico and Saint Thomas in the upper peneplane. (2) The even hill snnimits of eastern Porto Kico, Culebra and Vieques may represent a peneplane contemporaneous with that of west- ern Porto [?ieo, but lower because it has been pushed back headward toward the Luquillo mountain mass by sliort streams operating closer to the sea than the rivers west of the monntain mass. If the spurs on the eastern flanks of the mountains represent the upper peneplane, as was conjectured above, tliis ex})lanation cannot hold. It appears unlikely, also, in \iew of tlie fact that the elevation of the Saint John peneplane accords rather exactly with the eastern projection of the tilted upper peneplane of Porto Eico. (H) The Vieques-Culebra hills, together with those in the Naguabo- Fajardo area, are merely monadnocks rising above the lower peneplane. According to this inter[)retation they are not referable to an inter- mediate baselevel, but attain subequal altitudes because of the uniformity of fluvial erosion acting on more or less homogeneous materials. The arguments against tins view must be discussed largely in connection with the factors favoring tlie fourth hypothesis. But the distribution of the even-topped hills, often without reference to drainage lines or distance from the sea ; and the old character of the surface where the inter- mediate elevations occupy broad areas, in strong contrast to the relative youth of more recent dissection, are two strong points against the view that regularity of skyline is an accidental development. Forms pro- (hiced in a single cycle of erosion have simply graded slopes and sliow^ uniformity in development. Youthful dissection of an T)ld surface indicates rejuvenation. (4) The Culebra- Vieques hills represent an erosion cycle in which the baselevel was intermediate between the upper and low-er peneplanes. The probability of an intertnediate or second erosion cycle is favored by several lines of field evidence: (a) Except for the marginal dissec- tion and the pronounced lowland between Great Harbor and Flamingo Bay, most of Culebra rises to a fiat upland, 300 to 400 feet high. Com- pared with the character of the upland surface the recent dissection is young, except in the Great Harbor-Flamingo Bay depression, (b) In Vieques tlie postulated level is best represented in the hill and ridge crests in the western half of the island, where the rocks are granitoid intrusives. Normally, the coarse-grained intrusives are reduced to low- lands with relative ease, and there are no peculiar features in the Vieques quartz-diorite wdiich make it resistant to erosion. The fact that this MEYERHOFF, PHYSIOdRAPHY OF VIRinX ISLANDS 191 area of intrusive rock .forms one of the largest continuous expanses of the erosion level under consideration and contains, besides^ the highest monadnock east of the Luquillo Mountai]is foot-hill helt (Mount Pirata) does not necessaril}- imply greater resistance to erosion, but may perhaps be attributable to a greater initial mass to be reduced and to a location remote from drainage lines. The intermediate upland surface formed upon this rock type is smooth and rolling ; the small valleys which are cut into it, and which are working headward toward the drainage divide, are young and steep-sided. In eastern A'ieqiies Avhere tlie island is narrow^ subsequent dissection has lowered the surface generally below 250 feet^ Mount Jalobre alone rising above 400 feet. Greater erosion in this part of the island was facilitated by its narrowness and by the presence of limestones interbedded with tlie volcanics. (c) The distri- bution of hills of similar height in western Vieques appears to be more regular and widespread than a haphazard develo])ment of monadnocks v/ith subequal elevations above the lower ])ene])laiu^ would render pos- sible. The same statement holds for the hills and ridges on different kinds of rocks in eastern Porto Pico, (d) The existence in the northern A'irgin Islands*"^^ of an intermediate erosion level of restricted develop- ment, w^hich is present in Savana Island (Part T, Fig. 28), only seven miles west of Culebra, leads to the expectation of a parallel erosional stage in Vieques and Culebra. Persistent and widespread areas in eastern Porto Rico at a similar altitude render inevitable the conclusion that this stage in the physiographic history of the Porto Kico-Virgin Islands bank was general and is of greater importance than has hitherto been recognized. The extent of its development in Porto Kico has not yet been determined beyond the Fajardo District, but presumably its development would be least where conditions least favored dissection of the upper peneplane or most favored the formation of the lower, or third-cycle, peneplane. Probably a restudy of the areas termed by Lobeck the "Rugged Foothill Belt/' as well as of the Tiower Peneplane, '^^ will further extend its boundaries. From several lines of evidence the conclusion seems warranted that the region between Porto Rico and tlie Virgin Islands has been sus- ceptible to rapid and extensive erosion throughout its history. It was one of the first areas to be reduced during the first cycle of erosion ; and upon renewed uplift the upper pene])lane was destroyed, and a new surface of fluvial denudation was cut nearly 1000 feet below the first. Later the second-cycle surface was all hut com]detely destroyed, 3^' Ante. pp. 30-42. ^ Lobeck, A. K. : Op, cit,, Fig. 3. faclnj? pape 'M)'t 192 HCIESTIFIC PURVEY OF THE VIRGIN ISLANDS remnants of it beinir well preserved only in C^ulebra and Vieques, which presumably were farthest removed from most active fluvial erosion. Indication of such a location is supplied by the presence of residuals in both areas, which rise as monad nocks toward the elevation of the upper pene|)lane—Mount Pirata on Meques, and Mount Resaca on Culebra. No final hypothesis can be advanced to explain why the entire area should have been an easy })rey to erosion. As western Vieques is under- lain by a normally weak granitoid intrusive identical, or closely allied, with a similar intrusive in southeastern Porto Rico, the drowned re- entrant b(e4:Aveen Punta Yeguas, Playa de Xaguabo and Punta Arenas is probably underlain by quartz-diorite. This portion of the bank might thus be ex[)lained as a drowned lowland worn in easily eroded rocks. There is no basis for believing that a similar explanation holds for the l)ank north and east of Vieques, for the intrusive is not encountered again for nearly forty miles (in Buck Island, south of Saint Thomas). The northwest-southeast alignment of topographic features between Culebra and Magens Bay, Saint Thomas, has already been described and attributed to faults of snudl displacement. The distribution of these faults, which are most abundant in the portion of the bank now' most widely submerged, implies that structural weakness was in part re- sponsible for th(* deep erosion east of Porto Rico and nortli of Vieques. ( I ranted the non-resistance of the region to erosion, the present position of the island residuals is logical: Vieques, located near the southern margin of tlie bank l>ecause low |)recipitation developed it farthest from the dominant northern drainage; Culebra, located on fairly dense resist- ant rocks witli nearly horizontal structure, possibly marking the center of a synclinal trough, midway between two old major drainage systems. Tni-: Lowi:!,' Pexeplane The coastal-|)lain deposit^ of \'ieques rest upon a hilly surface of fluvial denudation, which has been correlated by Lobeck with the lower peneplane of Porto Rico (Fig. 51). Only its marginal and more rugged portions are noAv ex|)o^ed ^here the coastal plain has been removed, but it extends as a surface of low relief beneath the Tertiary and beneath Meques Sound, Virgin Passage, and the area north of Cordilleras Reefs and Culebra. It was formed as a result of uplift of the second-cycle peneplaiu^ still re])resented in the hills and uplands of Culebra and \^ieques. Again, as in the second erosion cycle, the entire area between the Luquillo Mountains and Saint Thomas was reduced rapidly toward the new baselevel, until only the masses forming the hills of eastern Porto Rico, Vieques and Culel)ra remained as monadnocks. The sur- MKVF.inHH-'f, t'HYSIijilh'M'HV OF YIUHlX ISLAM): face has \mm iHiried bfiit'Hth i'<>H>trtl-plaiii ileiKwits, whieli iwMi it> tu not oiilv OH \'i(Mjii('.H, bill i\\>^) ill \]\v entire t\X|KUis»' whii-h imw ,;oiu tuto tlie Hul.nuirioc iiliitforni. It i< itriis iiiipiKSsible U> It'sirii uiivthi about tile siirtaee iletail tliat, lia.s (linni^ l)t'a,riji_i>- ii|)«iii if> oritrin ; v ^isido from flu- rii.ux'MJ surface exposed en \'ieqiies and the deep ih'p'r n> |>« sion in ('Jreat: llarbur and Fhiiningo Kay, .several coJiside to tlie fhivirti nrio-iii of tht» surfad'. One id* the L-u-avst K'etJons redu<-ed diirin.u' this Ihird erusiuii eveh' was the lowland now I'orndnii' \'ieipH's Sound, whieh is, and was at the time nf its I'orniation, slieltered from wave attack. 'J'here is no means of estimating liow much nniru-iiial t>rnsi(i.u was aceom|jlished by wa^e aetiini; but de-nndati«in was ilond- jniid:ly Hnvial amj seems to have been jiceumplishi'd larm'lv hv a ramifv^ iii|ir river system witli headwaters in the laiqnilh> Mountains, in the hnv divide between western \'ie(pies and Enseiiada Hcmda, Porto IJi«-o. and between tddebra and \"ieqnes. The protet-ted position of most of this region from wave attaelc and its eontVtrmaiiee to tdie phin of a river system seem eonvineing evideru'c ol" its Hnvial ori<«dn. The narrow low- land sontb (d* \de«jnes, where stream development was iiniin()ortant, is procd' of the ineffeetiven.'ss n{ the >ea as a factor in forinin»f the snrface. The small eays surrounding Cnlehra are nmnadnocks. indicating a rngged t«)pography which conhl not have existed, much b-ss liave loriited, under the persistent attack of waves. The third evele of erosion was of nearly as great duration as the second, j»r*«hn-ing a gron|) of connecU'd lowlands and marginal valleys from 500 to 1)00 feet below tlic upper level of Cnleln'a and \d'e«pies, M I'M SVIHX'JIFKJ t^rJiVfiY OF THE Uh'dIX h'-iLAM.hS in till? \'irMi,ii IsIhjhI^. thoe lowhriid* Hcrvtnl as h l»risenifnt oiwii which llic 'rcrtiary coaHtjj] |tlain was «lt'fKwit(Hl, and arc tliiis <'(HiternporaTie()Vir^ witli the lower iiPiH-filarir- of Furto Uk-o. The C'oA'Stal Pl.\in The 'l\'rtiiiry coastal ])liiiii lias a very liniittMl rcprcscntatioji in tla- J'oiio Rico jslaiid gnmiK Xo Tertiary deposits are fuiiiid oji I'u'lehrH. althoUMl) they iiiMJoulitedly iiiiderlit- the (VrrrlilJeras Ilwfs to tlie west. The distribution of Tertiary liinestonen on X'ieqnes. liowever, is si.miili- i-iint of their funiier widesprearl oeeiirreiice over a larg-e part of thi- i-ntire sidainirine phitforni area. 'J'heir best present expres.siun is on the am- tral part of flie soiitlierri eoasl. from Forto Um] to Chiha Cav in .^^r"::-—,;:^f Kiisenaihi de la Cliiba (Fig. 51 ). The inner linnts of the coastal plain are marked by a line uf hays, inelndiii.ii- Knsenaila Soiiibe, r><n1u 3los(nnto. Porto Ferro and Fiisenaibi de la Chiha {Plate IV). All ar-' dj'seoiinei-ted bays oeeiipving the partly drowndl inner Inwlaml (Fv2\ 5-M and are eoinieeted witli the I'aribbiaui by narrow months, eiittine thron«:-h the en«'sta whirh rises abont 100 t'eet alunv sea-level. When originally depositx'd, dtie liim-stones abntted against the irre«-idar froiiT of th,e C'retaeeonn liills, resting npon the nneven Jlnvial smd'aei' of tlie fin'rd-eyele peneplane, from whieh th«'y hnxv been ))artly renujved. At present they are about l•^") to 1 10 (Vet thiek. and it k donlitfnl if they were mueh fhii-ker when fornn^d. It is therefore uidik(dy thai thev • •overeil nnieh more of the island at this point: they did not cover any portion of the seeond-iwele perieplane. Tertiary liniestunes and marls. 1(50 to 1T5 fcnt tliiek, eompose the MEYERHOFF, PHYSIOGRAPHY OF VIROIX JSLAKDS 195 east end of Vieques from Porto de Saliua and Saliiia del Siir to Piinta Este. The narrow isthmus between the two bays connecting the coastal plain with the mainland is the partly drowned inner lowland, whose surface is further covered by shallow lagoons. The somewhat greater thickness of the deposits appears to be the result of less stripping during recent erosion than occurred along the south shore. The third occurrence of Tertiary consists of a patch barely two acres in extent; yet of the three Tertiary areas it is tbe most significant. It is located on the north shore at Punta Caballo, midway between Isabel Segunda and Punta Arenas. The strata underlie a Ioav hill which projects from the general coastline sufhciently to form the only promi- nent headland between' Punta Mulas at Isabel Segunda and Punta Arenas at the western end of the island. It cannot be doubted that it is the sole surviving remnant of a large expanse of coastal plain which once covered all of Vieques Sound. The thickness represented at Punta Caballo is not more than fifty feet. There is no pronounced inner low- land, for the deposits rest against the oldland hills which front upon much of the northern coast. The latter is developed on the Cretaceous volcanics and locally upon semiconsolidated alluvial deposits of post- Oligocene age, and consists almost throughout its length of gentle north- ward slopes, cut by rejuvenated valleys. The slopes end at the shore- line in a wave-cut cliff 20 to 30 feet high. The coastal section appears to be the erosion scarp between the second- and third-cycle peneplanes. It formerly served as the inner valley-wall of an inner lowland, from which all of the coastal plain except the patch at Punta Caballo ha*s been stripped. Before dissection and removal of the cuesta, the alluvial deposits mentioned above were deposited, aiul they now front the coast because the cuesta has been reduced. The failure of the coastal plain to rise above the surface of Vieques Sound elsewhere is due largely to the amount of dissection it has undergone ; in part, perhaps, to a smaller amount of initial accumulation. The original extension of the coastal plain is merely a matter of con- jecture, yet the limits of conjecture are so narrow that a fairly convinc- ing conclusion can be reached. During the development of the third- cycle peneplane broad fluvial lowlands were formed in Vieques Passage, in Vieques Sound, in Virgin Passage and in the area north of Culebra and the Cordilleras Reefs, while a narrower one was formed south of Vieques. The relative breadth of the lowlands depended upon the structural and textural weaknesses of the underlying Cretaceous rocks, the rainfall distribution and the arrangement of drainage systems. After their formation thev were submers^ed to a level somewhat below 190 mJIENTIFlC SURVEY OF THE VIRGIN ISLANDS the surface of the second-cycle peneplane. Coastal plain deposition ensued; and^ if the deposition now in process on the submarine plat- form may be regarded as typical, accumnlation of material was greatest near the shoreline, where it consisted of coral-reef growth and bars; it was less within the reefs, and least on the offshore sides of the reefs.. It became proportionately less at greater distances from the shoreline, where most of the accumulations were composed of the lime-sand and mud derived from the reefs. In this manner a coastal plain is accumu- lating around Anegada today; and, judging from the distribution of coral heads in the Oligocene deposits in Saint (Voix, these seem to have formed in an analogous way. This method may be considered to apply to the Tertiary of Vieques and the entire bank between Porto Rico and the Virgin Islands. The present inner lowdands on Vieques, accord- ing to this interpretation, were sites of accumulation of clastic material whose source was the oldland on the one hand, the reefs and bars on the other. The amount of accumulation from both sources was probably little less than along the reefs. In the open water beyond the reef zone accumulation was probably much slower, and the ultimate thickness of the deposits, less. Subsequent uplift of the coastal plain led to its dis- section. The reef material, consisting of firmly rooted coral heads with interspersed calcium carbonate debris and cement, has proved most re- sistant to erosion and now forms most of the uneroded coastal plain on Vieques. The areas within the reefs and bars, composed of poorly cemented debris, were rapidly eroded into inner lowlands; the seaward areas of deposition, filled largely with uncemented lime-sand and mud,, were likewise rapidly denuded localities, initially of lower elevation. The present distribution of the Tertiary on Vieques is thus explained. That it has been wholly removed from the bank which is now submerged is not at all likely. It is assumed, therefore, that much of Vieques Sound, Virgin Passage, Vieques Passage and part of the northern section' of the bank are underlain by remnants of the coastal plain, partly veneered by subsequent deposits. Complete absence of erosion remnants of the coastal plain on the fringes of Culebra and on the adjacent islands^ can be explained as the outcome of more vigorous erosion in the northern drainage areas, where rainfall tends to be slightly heavier, and a some- what greater amount of submergence at the present time. Uplift and Dissectjox of the Coastal Plaix Some of the factors affecting the dissection of the coastal plain after its uplift w^ere discussed in the last section in considering its original extent. There remains for consideration the mode of dissection. The MhJYERHOFF, J^HYSSIOGKAPHY OF VIRaiy ll^LAXDS I97 amount of uplift was not less than '')5() or 400 feet, an amount repre- sented by the present depth of the phitform and tlie elevation of the <*oastal })lain de|)osits above sea-level. Most of the erosion wliieh oeenrred after n])lift was restrieted to tlie soft Tertiary marls, and in eonse- quenee the lowlands were re-exeavated, while a eom])aratively small amount of erosion ocH'urred in the harder roeks of the old land. The drainage established upon uplift was eonsecpient, the greater part of it having a general northerly direction, deflected, however, by the }>osition of Culebra, and modified by local variations in the contour of the surface. The easily eroded marls were rapidly but not com})letely reduced to new lowlands between Vieques, Porto Kieo and (^ulebra, while the more resistant reef limestones surrounding Vieques were cut by young valleys. Behind them the less resistant material was dissected by headward- vvorking subsequent streams, which have formed strong inner lowlands from Porto Peal to Ensenada de la Chi ha in southern ^^ieques (Figs. 51 and 52), and from Salina del Sur to Porto Salinas near the eastern end of the island. There were thus formed elongate inner lowlands, with young, narrow valleys as outlets through the retreating cuesta. As a result of these features submergence to the present level has formed tlie ])()uch-shaped harbors along the southern coast. The recency of the removal of the flanking coastal plain in other portions of Vieques is amply attested by the ])rese]U'e of gravel-clay alluvial deposits, many of which are now marginal along parts of the coast. They are especially well developed in the valley one and a half miles west of Esperanza and between Punta Caballo and Hesolucion. The materials have been derived almost wholly from waste from the old- land rocks. Where they have been formed by disintegration of the diorite in western Vieques, they consist of coarse arkoses, with subordi- nate amounts of hornl)lende, quartz and clay; where they have been derived from the andesitic volcanic rocks, as in eastern Vieques, they <'onsist largely of graywacke-clays. They fill the larger valleys of the oldland and form conspicuous accumulations in the inner lowland of south western Vieques, where their arkosic character and low clay con- tent make them an important water-bearing sand. They exhibit good stratification, yet in their mode of occurrence ap[)ear to be of simple fluvial origin. They are older than the drowning which is now a fea- ture of the coastline and are at present undergoing dissection. There is no evidence that greater submergence of the island occurred, and they cannot be exyilained as deposits made while the sea stood at a liigher level. They a|)])ear to have resulted from temporary aggradation dur- ing maximum emergence of Porto Pico and the Virgin Islands bank. 198 SCIENTIFIC SI R] EY OF THE VIRGIN ISLANDS which was aceompaiiied by rapid deimdation of the oldland cores. Ag- gradation took place in the h)wer portions of valleys cut into the old- land, and particularly in the inner lowlands. The marginal alluvial deposits now found along tlie north coast in western Vieques are believed to ha\e been accumulated in an inner lowland and have since been left exposed to coastal erosion by complete removal of the Oligocene lime- stones and marls. xV surprising proportion of the surface material of Vieques below the elevation of the intermediate peneplane consists of * alluvial deposits so formed, but in Culebra alluvium underlies only a small portion of the lower levels. Some was noted in the valley between the heads of Great Harbor and Flamingo Bay^ but much of this is com- paratively recent. Other deposits were observed at the heads of the embayments of the southeastern section of the island. None of these deposits are marginal, and there is no indication that they formed within the cuesta in an inner lowland. Probably the dissection of the coastal plain is not the outcome of a single stage of uplift and denudation, succeeded by drowning. There seems to have been initially a moderate amount of uplift, followed bv partial excavation of tlie old lowlands, youthful dissection of the margi- nal coastal plain, and formation of cuestas and inner lowlands where the nature of the Tertiary deposits permitted. A second maximum uplift inaugurated such active erosion upon the oldland monadnocks of Culebra and Vieques, that aggradation of the valleys and inner lowdands occurred and was succeeded ultimately by degradation of the entire platform. Later submergence has largely stopped erosion, except for the slow denudation in process on Vieques and Culebra, while deposi- tion is again going on in all of the submerged areas. The Submarine Platform Fluvial dissection of the coastal plain effected the re-excavation of the third-cycle lowlands, wliose submergence has produced the submarine platform. The subnu»rgence appears to have occurred in two stages. the second of which is responsible for the present level. The history of the Culebra-Vieques submarine platform is an integral part of the history of the northern Virgin Islands platform; and, as the latter's development has been fully treated in the second chapter of the present report, only the special features found in the Culebra-Vieques region requ i re d iscussion . Perhaps the most striking feature brought out by a study df the surface of this part of the bank (Part T, Plate II) is the absence of anv '^^ubmarine terraces," whose occurrence is described bv Vaughan in MEYI'JRHOFF, I'HYSIOaRAfrHY OF ^ IRdlN ISLANDS 1«J<) the adjoining Virgin Islands bank. Tliere are, it is true, a large num- ber of shoals whicli rise from depths ranging between eight and eighteeji fathoms, and which are now submerged to subequal depths, usually five or six fathoms. How far these shoals represent eom|)aratively recent growths of coral reefs or fiat-to])|)ed mesas of the dissected Tertiary coastal plain, it is impossible to determine. Aside from the shoals the bank appears to express simplicity and unity of development. Promi- nent submerged escarpments are few and are limited to the eastern ends of Vieques and Culebra, the northern side of Culebra ami the Cordilleras Reefs, with a subordinate scarp on the soutli side of the latter and low scarps on the flanks of the shoals. Here, also, as in the northern Virgin Islands, the margin of the bank is an abru|)t droj) to depths of several hundred fathoms both north and south. It is accentuated in most localities by a narrow drowned reef or by broader shoals rising above the platform at its brink, although along the northern edge the shoaling usually occurs one to four miles within tlie rim. The entire surface of the bank is modified extensively hy coral growtli and by sand and mud deposits, which make it difficult to restore the erosional surface developed by fluvial dissection of the coastal plain. In a general way, however, the original surface of denudation can be approximately determined. The floor of the platform is exceedingly level over large areas. Xearly 100 square miles of Vieques Sound have less than 30 feet of relief; and approximately 150 square miles of tlie northern part of the bank, lying between Icacos Cay and Virgin Passage, and between the northern edge of the bank and Cordilleras Reefs and Culebra, present gently sloping to flat surfaces. The slo])e is northward aiul varies from 0-15 feet per mile. These two flat expanses are separated by Cordilleras Reefs and Culebra; they merge into Virgin Passage on the east and terminate westward in the shallow water, broken by reefs, shoals and islands, off Porto Rico. There the irregularity of the surface, in marked contrast to the more extensive level stretclies to the east, l>espeaks im- ])erfect dissection, complicated by coral growth on the drowned hilltops. The level area of central and eastern Aleques Sound slopes gradually from a depth of 8 to 12 fathoms off Vieques and Porto Rico to depths of IG to 17 fathoms in the central and northern portions of the sound inside Cordilleras Reefs. The deepest areas are basinal because of the irregular deposition of clastic materials on the bottom in recent time. Although a young marine plain of deposition is 1)eing formed, the flat- ness of the surface is not due to the accumulation of sediments, hut is rather the outcome of the earlier dissection of the liorizontal coastal plain. Ample proof of such an origin for the sound is afforded by the :>nii sriKXTific sch'vf-:Y of Till-: viiniis wi.axds bclmvior of* siibmariiie coutorii's at itn moiitli between onslfni \'ieijiii's iuid crtHteni CiiJclira. Depths uf 85 to JMi feet |)orHist alirH>>t t<. llu; outlet uf the souml, but <'ii<l against a low subiiiera'<l ridge wliich rises ^.ii t<. m feet abovi' the siirroutidiiig snrfaee. The ri.Ige is Jong j,ihI narrow, exti'ii.liiig Fruiii the Tertiary eri.-sta ou e.astcrn \'ie(iiie< to the shoals ertsi t)t' Ciilet.rn. The rulge is hn.kni by a pip, 84 feet der.p, three atid a quartiT mile> north <.f the coastal plain <»ii Meques, anil by a brosbler pi\K HO feet (le.'p. two niile« froni t'ulebra; bnt elsewlier*' it is re,n~ tiniH.iis. South of the smiiherii hrertk, the western slope, though gentle, is steeper than the eastern. The fonii and outline of this ridge, as widl as its position find its (•(uuHH-litni with the eoa^stal plain on, tnistern X'ieques, strongly imply that it is part of tlie Tertiary eoastal |dain. low euesta fa»'ing inward upon the \'ieques Soojid lowhmd. The gap near the eenti-r is an old valley whieh was eut through it in tlie course itf its diKse«-tion. Tiie eha/raeterislies to be expected of a dissected e-oastal plain are also found immediately ea.st oi the ridge and southeast of Cnlebra in (I'rampus Sboal.^ and tlubr prolongations. The nature vt Cordilb-ras Reefs, the harrier between X'ieques Sound' and the northern lowhin.l, (lenunnls earel'id consideraticm. It i-onsists of a long, narrow ridge with faint beginnings three and a half niihs n.u'th of liuqnillo. Porto llieo. attaining excellent devehq.ment one mile north of* Cape San .hum and continuing to Taetus Cav, three and a half miles west of Cnlebra, a total distance of twenty milrs. Ii rises above soa-Ievel in Las Cuearacdias, i.o^ Farallones. ' Jeac-os. Ihitones. liobos ami Ihahlo cays, and in ITermanos and Barriles reefs, as well sh in other iliminutive nnnanied islainls (Fig. o;]). The eastern half of tlie ridges rises from, a basr 80 to 1tMi feet ludi.w sea-level tti a Hat^foi.ncd Fio, M.—S'«rnrrn Vieqvm Smm^l antJ the mniern lintf of rnrtliUrrm Rrrp, Mftnmnfi the mthmnrme tnpoffrnphv MEYERHOFF, PHYSIOGRAPHY OF TlRdlN ISLANDS 0()1 summit averaging 30 feet below sea-level, but it attains the surfaee in numerons small rocks in Barriles and Hermanos reefs (Fig. 54). Along many miles of its length it is a double ridge whose northern and sonthern members are separated by a trongh of slightly dee]>er water. When it is double^ one ridge (commonly the southern) is strong; the other, weak and broken. In the eastern half of its extent the entire ridge rises from the platform base in a moderately steep escarpment on the north, and in an almost equally prominent escarpment on the south ; but in its western portion the southern escarpment is low and gentle, sinking gradually to the shallow depths (less than 11 fathoms) off the Porto Rico mai]dand. Bast of Blake Shoal the southern escarpment is uni- formly simple. On the north the scarp is simple only in its eastern portion (the Barriles Reef section)^ where it drops from five fathoms at the crest of the reef to seventeen fathoms. From Hermanos Passage westward, it changes to a low com|)ound slope with an average declivity of one degree. In the vicinity of Hermanos Beef and Diablo C^ay it slopes evenly and gently north-northwestward to a depth of 12 to 14 fathoms; at this level the declivity increases to two degrees to a depth of 21 fathoms; then flattens out to a slope of five feet to the mile toward tlie northern margin of the platform. The crest of the ridge varies betw^een a half mile and a mile in breadth; the base, from oiie to two miles (Fig. 54). The use of the name "reef" for the Cordilleras ridge is misleading. It suggests coral-reef origin, and although some coral heads were observed in active growth at scattered localities along its crest, they appear to have played but little part in its origin. Examination was made of the rocks com])osing the islands which rise above the reef. Although it is believed that these rocks do not compose the ridge to its base, they suggest some strong analogies to features found in the north- western part of the Fajardo District of Porto Rico, where origins and geological history were plainly decipherable. Icacos, Ratones, Lobos and Diablo cays are made up of consolidated thin- to medium-bedded lime sandstones or calcarenytes, which exhibit strong cross-bedding. The perfection of sizing and rounding of the grains, as well as their purity, imphes aH)lian origin, which, however, the irregularity of the cross- bedding does not fully confirm. The white color of the formation and the irregularity of its cross-bedding make it stand in strong contrast to tlie San Juan formation of Porto Eico, which is buff-colored and regularly cross-bedded. Yet in the peculiarity of its position, degree of consolidation, and perfection of sorting it resembles the San Juan forma- tion so closely that it will be considered an equivalent formation of 202 SCIENTIFIC HUllVEY OF THE VIRGIN UPLANDS wind-blown origin. From the data available, contemporaneity of origin may be assumed for both formations. A similar conclusion concerning their age and origin has been reached by Lobeck.^"^ In northern Porto Rico the San Juan formation is invariably asso- ciated with the coastal plain. In the Arecibo District the association is direct; but in the northeastern San Juan and northwestern Fajardo districts the formation—a long ridge paralleling the coast—rises higher than the coastal plain on which it rests, and so stands above sea-level, whereas the underlying coastal plain deposits have been submerged. The San Juan formation thus serves as a barrier behind which are embayments and lagoons. The easternmost outcrop of the San Juan formation is at Punta Vacia Talega, about two miles west-northwest of Loiza, although the coastal plain is probably buried beneath the surface at least as far as the mouth of Rio Grande. If the San Juan ridge on Punta Vacia Talega and the more westerly peninsulas on which it out- crops be projected eastward, the line of projection is found to coincide with the Cordilleras Eeefs. This suggestive alignment, worth little if considered by itself, acquires significance when taken with the other evidence that implies the equivalency of the San Juan formation of Porto Rico and the rocks of the CordiUeras cays. Uplift and dissection of the coastal plain took place in Miocene time; the age of the San Juan formation is considered tentatively as Pleisto- cene, and its origin is associated with resubmergence of the coastal plain after its dissection. A question immediately arises : Were Vieques Sound and the northern platform excavated as continuous lowlands during Miocene and Pliocene time? If so, or if not, what is the rela- tion of the Cordilleras Reefs calcarenyte to the coastal plain ? In the city of San Juan, its type locality, the San Juan formation has a thickness of not less than 100 feet, but in other localities a thick- ness of 50 feet is the average. It is doubtful whether any of the Cor- dilleras Beefs formation is more than 50 feet thick, but the entire ridge rises 100 feet above the platform. It appears likely, therefore, that it rises above a base of Tertiary material. A study of the submarine con- tours with a view to establishing the late Tertiary drainage lines u|)o^ the platform indicates that the Cordilleras Reefs barrier existed through the late Tertiary period of dissection, although subsequent deposi- tion upon the surface of the bank makes accurate interpretation impossible. Drainage from western Vieques, Vieques Passage and eastern Porto Rico (south of Rio Fajardo) seems to have been directed northeastward and eastward across Vieques Sound lowland, ])assing out 8^ Lobeek, A. K. : Op. cit.. pp. 348 and 302. MEYHRHOFF. rHYtilOG RAPHY OF VlRaJX JSLAXDS 208 upon the northern lowland through Tuna Passage valley, three miles west of Culebra, where superposition is suggested (Fig. 54). The most deeply dissected pot-tion of the Vieques Sound lowland lies in the nortli, sometimes at the foot of Cordilleras ridge, as might be expected if tlie ridge is a cuesta formed upon coastal-plain materials with a gentle dip to the north, liio Fajardo may have Joined this drainage system on the nortli, although the present depths suggest a northwest course for it througli San Juan Passage. -The drainage of Culehra, eastern Meques and the eastern lowland seems to have found an outlet througli the low ridge between the eastern ends of the two islands, making its way south- eastward to the Caribbean. In Virgin Passage drainage appears to have been north and south, the divide extending from the southern margin of Culebrita slightly north of east to Savana Island. A spur extended southeast from the main divide to Sail Pock. In western Vieques Sound several of the shoals, such as those between Hodgkins and Blake shoals, a])pear to represent not recent coral accumulation, but incom- pletely reduced remnants of the horizontal coastal-|)lain sediments. To what extent the other shoals may be of similar origin is not determinable from the data at hand, but miglit be learned from field examination of the rocks wliich protrude above the surface of the sound. Tlie northern and Virgin Pas-age areas of the platform have a simple surface, ])resumably erosional, but modified by deposits which have filled in most of the drainage irregularities. Their average level is slightly lower than that of the Vieques Sound lowland, as might be ex])ected from their proximity and direct relation to the late Tertiary baselevel. Yet the marginal ridges along their outer limits near the edge of the l)ank need consideration. The simplest interpretation of their origin is, of course, that they are barrier reefs, later in origin than the surface of the bank, and consisting of reefs that grew during late Pliocene or early Pleistocene submergence. Exposure or drowning terminated their development. This explanation seems applicable to the marginal ridge south of Vieques, where the narrow ridge encloses undrained basins, whose origin could not have been a normal feature in fluvial dissection. The series of ridges near the edge of the bank north of Cordilleras Peefs, Culehra and Virgin Passage offer a striking ccmtrast in several resotcts to the reefs along the southern edge of the bank. The latter hui( the margin of the platform and are continuous with its outer slope. They are narrow, rarely exceeding 300 yards in width, and commonly mfasuring less than 200 yards; and they ape every bend and angle of the platform boundary. They are amazingly persistent, extending more than 30 miles «outh of the Virgin Islands without a single gap cutting 204 HCIENTfFia SURVEY OF THE VIRGIN ISLANDS through to the base of the reef^, although there is moderate variation in the depth of its crest below sea-level. (These statements do not apply to the broken ridges south- of Virgin Passage.) To the regularity and sim|)licity of the southern reefs, the ridges near the northern margin present the following contrasts : (1) They are situated at varying distances inside the edge of the bank, rarely less than a half mile, commonly as much as two miles. (2) Their northern slopes are not continuous with the steeper outer slope of the platform, but have a declivity of less than two degrees. (3) Their trend conforms imperfectly with the trend of the platform margiji, and most of the local flexures in each are independent. (4) The northern ridges are uniformly broader than the southern reefs, a width of one mile being common; they are also more complex, exhibiting a variety of spurs and re-entrants. (5) They are not persistent. Nowdiere do they extend more than ten miles without a gap cutting through to their base, and shallower gaps are of common occurrence. Several similar ridges are found four to ten miles south of the edge of the platform north of Virgin Passage, while south of Virgin Passage three analogous isolated ridges or mesas occur along the border of the bank. The ridge three and a quarter miles east of Vieques is most prominent. From the contrasts enumerated it seems reasonable to conclude that a distinction must be made between marginal reefs and marginal ridges. The latter appear to represent partly dissected remnants of the coastal |)lain, whereas the reefs represent post-Tertiary accumulations of coral growth. This tentative explanation is offered for whatever it may be worth, to explain the strong divergence in the two types of marginal barriers. (Hilebra and its dependent cays occupy an isolated position, standing in strong relief above the surrounding submarine platform. On the west and south, the depth of the bank is 12 to 15 fathoms; on the east, it is 12 to 25 fathoms^ increasing regularly from south to north. Xorth of the island there is, in succession, a descent of 12 fathoms to a narroif trough parallel to the coast, a ridge rising to a maximum of 5^ fathoms also paralleling the coast a mile offshore, a low cliff 30 feet high facing the Atlantic, then a gentle northward slope to 13 fathoms; finally, a moderate to strong descent to the deeper platform level, which is (,iere 26 to 28 fathoms deep. The isolated position of Culebra seems to be the result of three factors : (1) the strength of the rocks composing it; (2) the structure of the MEYJJRHOFF, PHYSIOGRAPHY OF VIRGIX ISLANDS 205 rocks which lie nearly horizontal, and which })ro])ahly form a shallow synclinal trougli or basin ;^'' (8) an assumed favorahle position with regard to late Tertiary drainage lines. The alignment of cays on the western side of the island, such as the Cactus-Luispena series and the Fungy Bowl-Xorthwest Point chain (both independent of the north- west fracture system), seems to depend upon structure, but dilhculty of access to the outer cays made it impossible to verify this assumption. The shoals on the northern and eastern flanks of Culehra may he inter- preted in two ways: (1) They may be developed upon Cretaceous rocks dipping inward toward the axis of the syncline, with erosion scarps facing outward. Xorth of the island the absence of a strong inward- facing scarp and the comparatively steej) northward-facing scarp helow the 13-fathom level favor this interpretation. {2) They may he dis- sected Tertiary coastal-plain deposits—a view favored by the steej) inward-facing scarp bounding Culebrita Reef on Soutli Cliaunel, and by the apparent continuity of the slioals east and soutlieast of Culehra with the coastal-plain ridge extending southward to Meques. In either case the concentric arrangement of shoreline, channels and shoals is satisfactorily accounted for. Theoretically, the criteria on which a decision between the alternatives could be made are so shar[)ly defined that a ready conclusion might he expected from examination of tlie charts. Decision is rendered difficult, however, l)y the inadequacy of information contained on tlie small-scale charts. The conclusion suggested by the available data is that north of Culebra, where the domi- nant scarp faces away from the island, the shoal is formed upon Cre- taceous volcanics ; that east and southeast, where the most prominent scarp faces inward and the lower outward-facing scar|) aj)pears to depend largely upon accentuation by modern coral growth, the shoals are devel- oped upon submerged coastal-plain deposits. All considerations indicate that the development of the submarine platform from Porto Rico to Virgin Passage was the outcome of uplift of not less than 350 feet, following deposition of the coastal plain. The dominant factor in the formation of the platform was fluvial dissection, operating effectively upon the newly deposited coastal plain, less effec- tively and more slowly u|X)n the harder and remoter oldland rocks. As a ^ The conclusion that the structure of Coleln'a is syncUnal Is put forth tentatively, because It could not be completely verified by field observations. The rocks are essen- tially massive without observable dip or strike, except on the shoreline southeast of Flamingo Point, where gently undulating stratified volcanics showed no constancy in amount or direction of dip. The conclusion is based on the presence of a westward- dipping Cretaceous cuesta in the eastern part of the island, and the semicircular arrange- ment of islands, shoals, channels and escarpments from the northern to the southeastern side of the island. 206 HCIESTIFIC SURVEY OF THE VIRGIN ISLANDS rough estimate^ at least nine-tenths of the bank now under water was cov- ered by the coastal plain, and it is believed that Tertiary limestones, much reduced in thickness, still cover a large portion of the platform floor. There is no conclusive evidence that marine planation played any part ex- cept on the outer mile or two of the northern edge of the bank. If it could be proved that the marginal ridges are of comparatively recent reef origin, there would be grounds for assuming that marine planation might have been effective as far south as the Cordilleras Reefs and Culebra escarp- ' ments, but this possibility is not demonstrable. A series of profiles across the bank from south to north shows three apparently recurrent levels or terraces—an upper one at 4 to 6 fathoms, an intermediate one at 12 to IT fathoms (Vieques Soinid), a lower one at 23 to 30 fathoms (Part I, Plate II). The upper one appears to be of late origin and will be dis- cussed in the next section; the other two have been interpreted by A^aughan as the result of two distinct periods of marine planation. The present study has attempted to disprove the view that the platform was developed iji two cycles, and to prove that it was formed during a single cycle in which marine planation was of no importance. The plat- form floor has l)een modified to a considerable extent by post-Tertiary deposition, especially in Vieques Sound, where land-derived sediments, coral growtli and coral fragments, without means of egress to the more open portions of the bank, have been actively altering and building up the surface. Less extensive but important modifications have been made in Virgin Passage and on the bank north of Cordilleras Reefs and Culebra. The genesis of the platform appears to have jDeen in no way related to depression of the ocean level during Pleistocene glaciation. Vaughan in his study of this area^^ seems not to have taken into account the pre- Pleistocene history of the platform. His paper is primarily concerned with the so-called outer flats, particularly those north of Cordilleras Reefs and Culebra, and the corresponding ones north of Saint Thomas, Saint John and Tortola. According to his interpretation a submarine flat was formed at de])ths of IT to 20 fathoms, followed by emergence and wave planation of 'the platform at depths of 26 to 34 fathoms along its seaward margins, succeeded by submergence to the 6- to 10-fathom level. His evidence has already been outlined in considering the development of the platform in tlie northern A'irgin Islands. He says in part: ^'"The escarpment extending from the islands north of Culebra Island, east of Porto Rico, across the Virgin Passage, and along the north side of Saint Tliomas, and the escarpment on the face of the ^ Vaugban, T. W. : "Some littoral and siil)nttoraI physiographic features of the Vir- gin and northern Leeward Islands and their V>enrlng on the coral reef problem." .T(»ur. Wash. Acad. Sci., Vol. VI, 1010, No. a, pp. ri.'l-OO. MEYEREOFF, PHYSIOGRAPHY OF YIROiy JSLANDS 207 outlying shoals can be explained in no other way f namely, than as a wave-cut cliff at the landward margin of a submarine ])lain^ "developed when sea-level was about 20 fatlioms, or slightly more, lower than now/^*^ Against this assertion it must be contended that the escarp- ment does not exist in Virgin Passage: that it exists uorth of Saint Thomas only off islands and promontories—uever between them and that such an escarpment would have developed north of Culebra and Cordilleras Keefs by normal fluvial erosion because of the structure and strength of the rocks and their position with respect to drainage lines during erosion. In describing the platform north of the Virgin Islands, Vaughan ascribes to Barracouta, Kingtish and Whale banks an origin other than that of coral growth/^ Tliese banks or shoals resemble the marginal ridges north of Cordilleras Reefs and Culebra in con- tour and in their relation to the platform. All of tlicm, it is thought, are remnants of the Tertiary coastal plai]i. With these serving as buffers to the sea, manifestly the sea could do little work toward developing a sea-cliff behind them. Marine origin, therefore, for any portion of the platform except the narrow and irregular bench outside (north of) the marginal ridges appears unlikely. The formation of the platform lowlands was accomplished mostly before Pleistocene time, subsequent to formation of the coastal plain in Oligocence time and its emergence shortly thereafter. Submergence to the six- to ten-fathom level may have taken place before the Pleistocene, ])ut the migratory nature of the strandline during five stages of withdrawal and return of the water, corresponding to the five glacial advances, makes very doubtful any assumption that extensive marine planation was accomplished at definite levels during Pleistocene time. Submergence of the SriniAiaxE Platform The embayed shorelines of C^ulebra, Vieques and eastern Porto Rico, as well as the relations of the smaller islands to the coasts, are ample proof of submergence. Whetlier the snbnu^rgence is the result of a single change from the level at which the platform was formed, or of more than one shift in level, is the chief problem connex-ted with this stage of physiographic history. Two lines of evidence may be used to reach a conclusion : marine erosional levels and depositional leveli^. The former may be expected on the seaward margiii of the land areas, especially those facing the Atlantic ; the latter may l)e looked for in the coral reefs whose growth was connected with the submergence of the platform, and also in protected areas where clastic materials have accumulated. « Vaughan, T. W. : Op. cit.. pp. 50-00. " Vaughan, T. W. : Op. cit.. p. 59. 208 iSCnJNTJFlC SURVEY OF THE XIROIN ISLANDS One mile north of Ciilebra, and separated from it by a channel 12 to 13 fathoms deep, i^ a rocky bank which joins Northeast Cay and Cule- brita to the east and southeast, and which, because of this relation, was considered above to consist of Cretaceous volcanic rocks. Its flat crest ranges only from 51/2 to ? fathoms, bnt descends to 9 fathoms in the channel which separate-; the ridge from Northeast C^ay. It slopes gently, but not quite exenly, northward to a deptli of 11 to 13 fathoms, where there occnirs a distinctly sharper escarpment to 26 or 28 fathoms. A corresponding gentle nortlnvard slope from 514 to 11 fathoms is found north of Northeast Cay, and a flat is developed between 4 and 12 fathoms between C^ilebrita and PalacUi cays. Where Cordilleras Reefs have not been breached and are not modi fled by recent coral growth, their summit is uniformly level at 5 to 6 fathoms (Fig. 54), Where islands are present, as at Diablo Cay or at Northeast (^ay, the descent from sea- level to 6 fathoms is very abrupt and has all tbe characteristics of a wave-cut cliff'. The h>w nortbern slo])e of (Cordilleras Reefs is presum- ably due to the dip of tbe coastal-] )lain sediments which compose it, as discussed in the ])revious section. But the slope may have been modified or perfected by marine |)lanation. Marine erosion forms a w^ave-cut cliff approximately at sea-le\'el and forms a graded wave-cut terrace below . sea-level. The {)resence of these features 25 to 30 feet below sea-level indicates a stand of the sea at tliat elevation sufficiently long to plane off platforms as mucli as a mile across. The level is recurrent in many localities, cliief of whicli are numerous shoals (including Hodg- kir.s) in the western part of Vieques Sound and local stretches along southern Vieques. If baselevel rem lined at tbe 5- or 5l/^-fathom level long enough for tbe |)lanation of terraces one mile wide, considerable filling of embay- ments must have been accomiilished. Because dissection had proceeded to a much lower biselevel j)rior to submergence, fairly deep embayment of the larger valleys occurred, and alluvial filling followed when these were drowned. Deposition durin/z this stage of development must have .been far more extensixe tlian that which foUow^ed upon the final sul)- mergence, which occurred so recently that comparatively few of the normal effects have had op|)ortunity to develop.*^ It follows, then, that the depth of water in many of the embayments will offer testimony for or against a baselevel at 5 or 5I/2 fathoms. Thus, in eastern Vieques, Ensenada Honda encloses two depressions, one at the eastern side 4 to fathoms deep, connected with the sea by channels less than a fathom ^^Tompare the width of tho wave-cit terrace formed during the final submergence, where recent sHght uplift has exposed it : I.obecl«, A. K.. op. clt., Fig. 37, p. 367 ; Fig 38. p. 360 : .Tohnson. D. W., SlioreUne processes and slioreline development, frontispiece MEYERHOFF, PHYSlOGRAFHr OF VlKaiK INLANDS 2()i) deep ; the other on the western side, 4 to 4% fathoms deep, witli a narrow outlet of 314 fathoms. Salina del Snr is 314 to 4U_> fathoms deep, with a slender channel connection of equal depth to the southeast. In Culehra most of Great Harbor shows soundings of 4 to 7 fathoms, with water 13 fathoms deep immediately outside the narrow mouth. Fla- mingo Bay, almost completely enclosed by shallow water, has depths of 4 fathoms in much of its center. South Channel affords a very striking case of partial filling: south of Culebrita a narrow belt nearly one mile long is consistently 12 to 11 fathoms deep, while the outlets to the nortli and south have been filled to depths of 6 or 7 fathoms. Other examples might be taken from eastern Porto Rico {e. g., Ensenada Honda) ; but those given suffice to show that in the larger bays and channels which were cut deeper than 5 fathoms during late Tertiary dissection, sub- mergence was accompanied by alluvial filling, which tended to build up their bottoms to a depth of 4 to 6 fathoms. All of the smaller bays and parts of the larger ones have been built up higher than this depth during the greater submergence which followed, but the evidence of the existeiu»e of a lower base of deposition has not been fully destroyed. South of Vieques there is a narrow drowned reef at the margin of tlie platform, analogous in all respects except in length and regularity to that described south of the British Virgin Islands and Saint John. Tlie reef was built up to 6 fathoms in places, although an average depth is 8 to 10. Its growth was apparently contingent upon submergence to the five-fathom level, and its death was probably not the result of sudden drowning to the present sea-level, but of exposure during tem|)orary withdrawal of the water during Pleistocene glaciation. From these three lines of evidence; namely, (1) marine planation, (2) alluvial filling of drowned valleys, and (3) growth of a marginal reef south of Vieques and Vieques Passage, all to a baselevel at an approximate depth of ^\e fathoms, it seems certain that submergence was accomplished in at least two stages, one to the level just described in preglacial time ; the other to the present level in postglacial time. A special phase of the earlier submergence was the formation of dune sands above high-tide level along the crest of the Tertiary cuesta which underlies Cordilleras Eeefs. In discussing the origin of the San Juan formation on the Cordilleras Reefs, Lobeck**^ relates its formation to the submergence which brought the strandline to its present position. So far as the Cordilleras Reefs occurrence of the San Juan formation is concerned, however, its origin must have antedated the latest sub- mergence. Similar sands are forming today above high water along ' Lobeck, A. K. : Op. cit., pp. 361-366. 310 SC/ENTIFW STJRYEY OF THE VJRGJN JSLANDS MEYERHOFF, PHYSIOGRAPHY OF VIRGIN ISLANDS 211 extensive stretches of sandy beach, often on a scale comparable with the extent of the San Juan formation. The dune sand forming the ridge along the northern shore of Anegada is an example which in all respects is analogous to the Pleistocene dune deposits of the Cordilleras Reefs. The Anegada sands are unconsolidated and have been derived from the clastation of coral heads and limestone in the adjacent waters. They rise above a comparatively recent section of the coastal plain and lie entirely above the level of high tide. The Cordilleras Reefs calcarenyte, it may be assumed, underwent similar development in the Pleistocene along the crest of the unsubmerged Tertiary cuesta. Its consolidation occurred with post-Pleistocene submergence, by which it was carried partly beneath the water-level. The thickness of the dune sand may be estimated as the sum of the elevation of the islands composed of it and the 25 to 30 feet which now lie below sea-level, a total of not more than 75 feet. It seems probable, also, that final removal of the Tertiary cuesta which surrounded Vieques, and behind which alluvial material had accumu- lated, was largely effected while the sea stood at the five-fathom level. The Present Level The final drowning of 30 or 33 feet, which brought the sea very nearly to its present level, produced few effects which have not already been encountered in the northern Virgin Islands. The filled, or partly filled, embayments again became areas of alluvium and sand accumulation, and most of them have been built up in part above the five- to six-fathom level of earlier deposition. The inner lowlands of southern and eastern Vieques were drowned, with the production of discontinuous flask- and bowl-shaped embayments (Fig. 55). These are sometimes all but con- nected by shallow lagoons, such as that between Porto Ferro and Porto Mosquito, and the two between Salina del Sur and Porto Salina. A few of the shallow bays have had bars formed across them, as at Playa Grande, Vieques, where a bay-mouth bar encloses a large lagoon, and at Flamingo Bay, Culebra, where a mid-bay bar forms the seaward side of a small lagoon. The lagoons in the vicinity of Punta Arenas appear to have formed as a result of bar formation by littoral currents moving northwest along the west coast and west along the north coast, with the spreading and piling of the sand by wind, and probably with slight broadening of the bars by very slight emergence. The converging lit- toral currents from the north and west shores of Vieques are building Punta Arenas northwestward into Arenas Bank, a lime-sand bank, one and a quarter miles wide at the base, tapering gradually to a slender 212 HVIENTIFW SURVEY OF THE VIRGJN ISLANDS point three miles northwest of its unsiibmerged extremity—Pmita Arenas (Plate IV). The development of Arenas Bank, bar formation and partial filling of the embayments are the chief depositional effects of the latest submergence; but others merit mention. Double tombolos connect the point between Porto Eeal and Ensenada Sombe with the mainland (Figs. 52 and 55), the northern and southern hills of Luis- pena Cay with the central block of the island, the Cape Passage penin- sula of Culebrita with the rest of the island, and Point Soldado with the peninsula forming the southwest side of Great Harbor. Upon numerous shoals in Vieques Sound, and around portions of Vieques and southwest Culebra, coral heads and small reefs are growing vigorously,, but elsewhere this activity is unimportant. Graded beaches have been developed around most of Vieques and southern Culebra, but on the north side of Culebra (except along Northwest peninsula) a beach is rarely present, depths of four to nine fathoms occurring close inshore. The filling of the drowned lowlands has, of course, continued, but there is no means of differentiating the results accomplished before and since maximum submergence. Some erosion has also been accomplished toward the present baselevel. The fluvial erosion cannot be distinguished from that of the preceding level, but the amount is presumably small. Marine erosion, however, has accomplished considerable work in central and western Vieques and on the Cretaceous headlands and Tertiary cliffs in the eastern portion of the island. Large areas of the island slope moderately toward the coast, but the slope is truncated by a wave-cut cliff, 15 to 30 feet high, which has been cut indiscriminately in Cretaceous volcanics and intru- sives, Tertiary limestone and late Tertiary or Pleistocene alluvium. The higher cliffs are usually associated with the limestones and alluvium, which, presumably, have been cut back farther because of their inferior resistance. It is a striking fact that none of the cliffs observed are now being actively worn back, but usually front upon narrow wave-cut ter- races, strewn with coarse debris. The terraces are commonly awash where narrow and formed upon rocks, but they extend beyond the wash of the waves where broad and developed upon alluvium. Hecent slight emergence has thus exposed the wave-work accomplished during maxi- mum drowning, so that the amount can be at^Hirately measured. The "vvave-cut terrace varies in width in direct proportion to the deirree of consolidation of the material upon which it was cut. On the alluvimn along the coast of northwestern Vieques it may be as much as 30 or 40- yards wide, though usually less. In the limestone areas along the soiitli shore it. is rarely more than four or five yards wide and is often absent,. MEYERHOFF, PHYSlOGRAPUr OF VIRdJX JSLAM)^"^ 213 particularly on headlands, where a steep, narrow beach may take its' place. In the Cretaceous intrusives and volcanics it is only rarely present as a distinct bench, but is replaced by a steep, narrow beach, grading upward to a wave-cut cliff beyond the reach of present wave- cutting. On Culebra, where only Cretaceous volcanics are present, there is little visible result of recent marine erosion. 8teep beaches occa- sionally occur, except along the nortli coast, where the deptli of water offshore has thus far prevented wave erosion. For this reason, also, evidence for recent slight emergence is meager. On the cays of Cor- ililleras Reefs wave-cutting at the level of maximum submergence ])ro- leeded vigorously and was apparently responsible in large part for the broad shallow-water areas, less than one fathom deep, surrounding several of the larger cays. The smaller oiu^s suffered most of their erosion at the pre-Pleistocene level, and characteristically drop off into five or six fathoms of water without any higher terrace developed u})on their flanks. In some cases, at least, their surfaces represent the marine terrace of maximum submergence. Recent Slight Emekm^enc e In the preceding section the fact was brouglit out that the X'ieques- Culebra area has undergone very recent uplift of not more than ten feet For convenience, the evidence will be briefly summarized without additional discussion. (1) Narrow wave-cut terraces, varying in -width with the type of material in which they were formed, occur locally above the wash of the waves. (2) Wave-cutting has ceased on practically all of the wave-cut cliffs (except in northern Culebra, where apparently it never began), which are now above the reach of the waves within high beaches or the terraces described above. (3) The height and declivity of many of the beaches fronting wave- cut cliffs indicate that they are but partly graded to the ])resent sea-level and were formed initially at a higher elevation. (4) Bars and tombolos, whose materials are unmodified l)y wind action, show a width and elevation that is not compatible with normal . development at the present w^ater-line. CONCLUSIONS Except for minor points of difference in the land-forms of the A'irgin Islands, Saint Croix and the Culebra-Vieques area, the physio- graphic history of all three areal divisions lias b(*cn essentially alike. 214 SCIENTIFIC f^lRVm' OF THE VIRGIN ISLANDS So close is their relation that tlieir physiography coukl have been treated under another areal grouping with equal propriety. Thus^ Culebra might have been iiiehided with the northern Virgin Islands, for, like them, it lacks an emergent coastal plain. Saint Croix and A^ieques, both of whicli are partly overlapped by Tertiary coastal-plain deposits that are absent on tlie other ishuids, are closely related not only physio- grapliically and geologically, but botanically as well. They could with good reason have been considered jointly, for their chief differences are largely the result of unequal amounts of coastal-plain dissection and submergence. It is not surprising to find a close relationship between the land-forms of Cide[)ra and Meques and those of the northern Virgin Islands. The fact that they all rise from the same shallow platform, directly connected with eastern Porto Rico, is in itself indication that their development has been closely parallel and has been similarly parallel with that of Porto Rico. That the land-forms of Saint Croix, notwith- standing the island's [)resent isolation, should also display the same succession of events not only is surprising, but has a significant bearing upon the tectonic history of this part of the West Indies. The charac- teristics and even the elevations of its land-forms are comparable in every respect with those of the islands to the north. iV review of the physiographic forms on all of the islands and their relation to those of Porto Rico will serve to emphasize the unity of their development. The Upper Cretaceous history ol vulcanism and subordinate sedimen- tation, terminated by orogeny and uplift, was the same from the western end of Porto Rico to A'irgin Gorda (Part I, Fig. 1), and from Tortoia and Jost Van Dyke to Saint Croix. A large, single oldland was formed, and during early Tertiary time it imderwent three cycles of erosion, the first two terminated by uplift, and the third by submergence. The first cycle formed the upper peneplane of Porto Rico and the Saint John peneplane of the northern Virgin Islands, leaving but few residual monadnock groups rising above the fluvial surface which was formed. The second cycle** caused the destruction of much of the earlier peneplane ** Since the present report went to press, the writer lias spent four months in Porto Kico, studying tlie geology of the Arecibo district. In that portion of Porto Rico a late Tertiary erosion-level wjts discovered beveling oldland rocks and coastal plain alike, and lying about SOO feet below the first-cycle peneplane. Its relationships led to the sus- f)icion TTiat it might be the equivalent of what has been considered the second-cycle erosional surface In this report, and subsequent field work changed the suspicion to a certainty. The Inte Tertiary level was not identified by Lobeck, and in the Arecilio dis- trict its extensive development at the expense of the first-cycle peneplane led him to confuse the two. It appears, therefore, thnt the physiographic history of the area as outlined In the present report must be revised to the following succession : 1. F^irst cycle—upper (Saint .Tohn) peneplane. 2. Second cycle—^lower (sub-coastal plain) peneplane. ,1. Coastal-plain deimsition. MEYERHOFF, PHYSIOGRAPHY OF VIRGIN ISLANDS 215 and produced an old erosional surface approximately 7^00 feet below the first. Thus, the upper peneplane was completely clestroyed in Saint Croix and in the Culebra-Vieques area, and the uplands of these islands represent the second cycle erosional surface. The third cycle formed the lower peneplane of Porto Rico and a mature to old surface which extends beneath the coastal plain on Saint Croix and A'ieques, and which under- lies remnants of the coastal plain on the submarine platform. Forma- tion of the lower peneplane was followed by deposition of a coastal plain in the middle Tertiary, and during late Tertiary time uplift ex])osed the coastal plain marls and limestones to dissection. Tlie Tertiary de- posits were formed in the entire area now constituting the submarine platforms of the islands, and the features of the i)lat form-floor can in many cases be identified as fluvial forms of a dissected coastal plain. Toward the close of the Tertiary differential movement or warping caused submergence of the eastern Porto Pico aiul Mrgin Islands regions, while western Porto Pico remained elevated. The submergence in- creased in magnitude as far east as Tortola, and there was formed a series of littoral features of marine erosion and deposition, wdiich have been submerged an additional five fathoms since the Pleistocene. N^ot only are all the physiographic events recorded in the land-forms of Porto Rico, Saint Croix and the northern islands, but the land-forms are spaced vertically at subequal distances. Thus, there is not a varia- tion of as much as 200 feet in the relations of first-, second-, and third- cycle erosional surfaces in any part of the region. Only the levels de- veloped by the two final stages of sul)mergence are out of line and fail to parallel the earlier land-forms. The upper peneplane descends from an elevation of 2,000 feet in Porto Rico to 1,000 feet in Saint John. Similarly, the second-cycle surface descends eastward a corresponding amount. The third cycle, or lower peneplane, and the base of the Ter- tiary descend from an elevation of 600 feet near Lares to more than 200 feet below sea-level off Tortola. In a similar manner, the second and third-cycle surfaces and the base of tlie coastal plain descend some 400 or 500 feet between Saint Croix and the northern Virgin Islands. On the other hand, the five-fathom level of wave erosion and marine accumulation is parallel to the present surface throughout all the islands 4. Uplift, wnrping nnd peneplanatlon of the coastnl plain. 5. T'plift, second cycle of coastal-plain dissection. 0. Siilimergence, as outlined on preceding pages. Tlie findings of tlie season's field work were presented before the Geological Society of America in December, 1020, and will appear in [)rint, it is hoped, in 1027. Further treatment of the subject will be published ultimately in the report on the geology of the Arecibo district in Volume 11, Part IV, of the Scientific Survey of Porto Rico and the Virgin Islands. 216 SCIENTIFIC SURVEY OF THE VIRGIN ISLANDS arid, like the present sea-level^ bevels the earlier land-forms indiscrinii- nately. From these relations it appears that the entire Porto Kico-Saint Croix- Virgin Islands area developed as a unit until the late Tertiary dissec- tion of the coastal plain was accomplished. Before the Pliocene- Pleistocene submergence, the entire region was tilted downward toward the northeast, and presumably at this time (Pliocene) the block-fauJting occurred which separated Saint Croix from the northern islands. The last, or maximum, submergence has affected the entire region alike only because it involved a general shift in the strandline, and not a differential change in elevation in any of the islands. BlBlVKXfKAPHY The excellent and full summaries which Professor Kemp has given in Part I of this vohinu^ of the literature bearing directly on the geolo<jy of the Virgin Islands, would appear to make further bibliographic material superfluous. In the text of "The Physiography of the Virgin Islands, Culebra and Vieques," moreover, quotations and citations of opinions from other authors have been accompanied by footnote references to tlieir works. It will be noted, however, that two-thirds of the references in the foot- notes are to works which have not been included in Professor Kenip\s bibliography, because they are not concerned directly with the geology of the Virgin Islands, C^ilebra and Vieques. It was inevitable that a study which involves not merely local physiography but correlation of land- forms over a wider geographic area, would be forced to call upon litera- ture dealing with a narrower scientific held, but a broader geographic field, l^'he references which have been used include only a few of those reviewed by Professor Kemp, but embrace many which had no plac^e in his summary. There has been, furthermore, no occasion to mention in footnotes many such works which were consulted, and which in all fair- ness merit mention. A bibliography appears, therefore, to be necessary, tlie more because it will assemble in a single section references to all the literature employed in the present physiographic study. Inasmuch as the physiographic history hinges upon the tectonic history of the Antilles, s(*veral titles dealing with this subject are included, even though they have not been helpful in any phase of the present work. The bibliography submitted thus includes: (1) Works dealing specifically with the geography and physiograj)hy of the Virgin Islands, (^ilebra and Vieques. MEYERHOFF, PHYSIOGRAPHY OF VlRdJX J8LANDS 217 (2) Works furnishing data concerning the geology of the Virgin Islands and Porto Rico that contributed to pliysiographic interpreta- tions. (3) Studies concerned with the physiography of Porto Eico and other Antillean islands, with which correlation has been made. (4) Papers or longer contributions pertaining to the tectonic history of the West Indies. (5) General geologic works, or articles dealing with specific areas other than the West Indies, from which analogies have been drawn. The titles are arranged alphabetically by author, without classifica- tion and without critical comment. For other works dealing with tlie geology of the Virgin Islands, the reader is referred to Professor Kemp's splendid summary of the literature iu Volume IV, Part I. Berkey, Charles F. 1915. (ieolo^ical reconiioissaiice of Porto liieo. Ann. N. Y. Aead. Sci., Vol. XXVI, pp. 1-70. 1019. Introduction to the (ieologj^ of I'orto Rico. Sei. Survey of Porto Kieo and the Virgin Islands, N. Y. Aead. Sci., Vol. I, pt. 1, pp. 11-28. Britton, N. L. IDltJ. The Vegetation of Anega<la. Menioii-s N. Y. Bot. (harden, Vol. VI, Pl>. 565-580. 1919. History of the Survey. Sci. Survey of Porto Hioo and the Virgin Islands, N. Y. Acad. Sci., Vol. I, pt. 1, i)p. 1-10. Daly, K. A. 1915. The glacial control theory of coral reefs. Proo. Amer. Acad. Arts and Sciences, Vol. LI. ])AVLS, W. M. 192(1 The Lesser Antilles. Ainer. (ieog. Soc. of New York. Fauoner, J. 1). 19(r2. Evolution of the Antilles. Scot. (Jeog. Mag., Vol. XVIII, pp. 300- 376. Fath, a. E. 1921. Origin of the faults, anticlines, and buried •granite ridge" of the northern part of the Mid-Continent oil and gas field. U. S. («. S., Professional I*aper 128, i>p. 53-06. Fettke, C. li. 1924. <ie(»logy of the Humacao District of l*orto Rico. Sci. Survey of Porto Rico and the Virgin Islands, N. Y. Acad. Sci., Vol. II, i)t. 2. <;re«ory, J. W. 1895. Contributions to the paleontology and pliysical geology of the West Indies. Quart. Jour. (ieol. Soc. of Londiai. V(»l. LI, jip. 255-310. 218 l^aiENTJFlC SURVEY OF THE VIRGIN ISLANDS GUPPY, R. J. L. 1909. Oeologieal coinieetions of the Caribbean region. Trans. Canadian Inst., pp. 373-391, January. 1911. On the geology of Antigua and the other West Indian islands with reference to the physical history of the Caribbean region. Quart. Jour. Geol. Soc. of London, Vol. LXVII, pt. 4, pp. 681-700. Hodge, E. T. 1920. (ieology of the Coamo-Guayama District of Porto Rico. Sci. Sur- vey of Porto Rico and the Virgin Islands, N. Y. Acad. Sci.. Vol. I, pt. 2. Hubbard, Bela. 1920. Tertiary formations of Porto Rico. Science, Vol. LI, pp. 395-396, April. 1923. The (Jeology of the Lares District, Porto Rico. Sci. Survey of Porto Rico and the Virgin Islands, N. Y. Acad. Sci., Vol. IT, pt. 1. Johnson, D. W. 1919. Shore Processes and Shoreline Development. John Wiley & Sons, Inc., New York. LlOHTBOURN, A. (i. 1923. Lightbourn's annual and commercial directory of the Virgin Islands of the United States. Saint Tliomas. LOBECK, A. K. 1919. Five land features of Porto Rico. Natural History, Vol. XIX, pp. 522-540. 1922. The Physiography of Porto Rico. Sci. Survey of Porto Rico and the Mrgin Islands, N. Y. Acad. Sci., Vol. I, pt. 4. MiTCHEU., G. J. 1922. Geology of the Ponce District of Porto Rico. Sci. Survey of Porto Rico and the Virgin Islands, N. Y. Acad. Sci., Vol. I, pt. 3. QuiN, J. T. 1907. The Building of an Island, being a sketch of the geological s.tru("- ture of the Danish West Indian Island of Saint Croix, or Santa Cruz. Published by the author in Christiansted, Saint Croix. Reeds, Chester A. 1919. New Base Map of Porto Rico. Sci. Survey of Porto Rico and tlie Virgin Islands, N. Y. Acad. Sci., Vol. I, pt. 1, pp. 30-31. SCHOMBUROK, ROBERT HERMANN. 1832. Remarks mi Anegada. Jour. Royal Geog. Soc. of London, ^'ol. 11, pp. 152-170. Semmes, D. R. 1919. (leology of the San Juan District, Porto Rico. Sci. Survey of Porto Rico and the Virgin Islands, N. Y. Acad. Sci., Vol. I. pt. 1. p[). 33-110. Spencer. J. W. 1894. Reconstruction of the Antillean continent. Bull. Geol. Soc. of America, Vol. VI, pp. 103-140. JIEYERHOFF, PHYSIOGRAPHY OF VIRGIN ISLANDS 9|(3 Spencer, J. \V. 1898. West Indian bridge. Pop. Sci. Mo., Vol. LlII, pp. 577-593. 1[K)2. The Windward Islands of the West Indies. Trans. Canadian In- stitute, Vol. VII, pp. 351-371. 1903. On the geological relationship of the volcanoes of the West Indies. Victoria Inst. Jour. Trans., Vol. XXXV, pp. 189-207. United States Coast and Geodetic Survey. 1921. Coast Pilot, Porto Rico and Virgin Islands. Dept. Commerce, Serial No. 164. Vaxghan, T. W. 1916. Some littoral and sublittoral physiographic features of the Virgin and northern Leeward islands and their bearing on the coral reef problem. Jour. Wash. Acad. Sci., Vol. VI, pp. 53-66. 1919. Some features of the Virgin Islands of the United States. Ann. Assoc. Amer. Geog., Vol. IX, pp. 78-82. 1923. Stratigraphy of the Virgin Islands of the United States and of Culebra and Vieques Islands, and notes on eastern Porto Rico. Jour. Wash. Acad. Sci., Vol. XIII, pp. 303-317. •'^ ^Mm&m^- m^-:;^ DO NOT CIRCIJI, :m ia :# '14'. PARTS ISSUED Volume I (eomplete in four parts, with index). Part 1. History of the Survey, by N, L. Britton. Geological Introduction, by 0. P. Borkey. Geology of the San Juan District, with colored map, by D. li. Seninies. Part 2. Geology of the Coanio-Guayania District, with colored map, by E. T. Hodge. Part 3. Geology of the Ponce District, with colored map, by G. J. Mitchell, Part 4. The Physiogra|)hy of Porto Pico, witli (M)]orcd map, by A. K. Lobeck. Volume II, Part 1. Geology of the Tjares District, witli colored map, by Bela Hubbard. Part 2. (jcologv of the Humacao Disirict, witli coloied ma|), by Charles P. Pettke. Volume III. Part 1. Tertiary Mollusca from Porto Pico, by C. J. Maury. Part 2. Tertiary Mollusca from the liares District, by Bela Hubbard. Volume: IV. (hM)logy of the Virgin Islands, (Aulchra and X'itMjUi-s. Part 1. Introduction and Pcview of the Literature, hy d. V. Kemp. Physiography, by H. A. MeyerholT. Part -2. l^hysiograpliv (coik hided ), by 11. A. MeyorliolT. Volume A' (com[)lcte in four ])arts, with gen(^ric index). Part 1. Descriptive Flora,—S[)ermatuphyta (Part), by N. D. Britton and l*ercy W'iWon. Part 2. Deseri})tive Flora— S[)(n'!natophyta (( 'ontirmed). Part 3. Descriptive Flora—Spermatophyta (Continued). Part 4. Descripiiv(» Flora— Spermatophyta (Coniinued). Volume VI. Part 1. Descriptive Flora—Spermatophyta ((Nnitinued) . Part 2. Descriptive Flora—Spi^rmatophyta (Continued). Volume IX. Part 1. (-hiroptera and Insectivora, Living and Extinct, by If. F. Anthony. Part 2, Hodentia and Edentata, by 11. E. Anthony. The above parts may be procured at $2 each by addressing Secretary, New York Academy of Sciences, 77th Street and Central Park, West, New York, N. Y. ^ .^i^fti- f 1^ ^^VoR< A<^AOeHV(iF^reyt^ lyK'y B 52802 4 ^ "'-^^W^- 3 9015 03458 5631