[AMERICAN JouRNAL OF Sc!El'\CE, Ym. 25,9, :\lARCH 1961, P. 211-222] THE PROBLEM OF REBEDDED POLLEN IN LATE-GLACIAL SEDIMENTS AT TAUNTON, MASSACHUSETTS ?\1ARGARET B. DAVIS"- California Institute of Technology, Pasadena, California ABSTRACT. The occurrence of pollen of tree genera not found in the modern flora of the United States and Canada in late-glacial sediments from the Williams Street bog near Tannton, :\[assachusetts, is considered evidence of contamination of the sediments with secondary pollen rcbeddcd from pre-Quaternary deposits. Nearby varved clays also contain these poll('.n types, and because the local till contains little if any pollen, it is concluded that the secmHiary pollen "-as carried in glacial meltwater from the vicinity of Tertiary deposits 25 miles to the northeast. The late-glacial bog sediments were apparently con taminated either directly by meltwater streams, or through the erosion and redeposition oi meltwater strPam deposits. As the secondary pollen spectrum cannot be determined with certainty, it is impossible to interpret the mixed secondary and primary pollen spectrum contained in the late-glacial sediments. Il\'TRODUCTION American pollen analysts have recently become concerned with a serious source of error in pollen analysis, the contamination of deposits with pollen from older sediments. Pollen grains are sometimes able to undergo more than one cycle of deposition without perceptible change in morphology; thus a single stratum may contain morphologically simiiar pollen grains derived from plants widely separated in time. unless the rebedded or "secondary" pol len is distinguished from "primary'' pollen ,derived hom vegetation living at the time the sediments were deposited, its occurrence in sediments may lead the pollen analyst to erroneous interpretations of vegetational history. The presence of rebedded pollen in late-glacial deposits in Denmark was demonstrated by I Yersen ( 1936). He found that boulder-cia y and varved clay in the vicinity of contaminated deposits conbained pollen and other microfossils from Tertiary sediments that had been over-ridden and eroded by the ice sheet. The late-glacial minerogenic sediments ·were contaminated with a similar as semblage of pollen but contained in addition pollen derived from the late glacial vegetation. Iversen showed that, if the spectrum of secondary pollen was determined from samples of nearby boulder-clay or van-ed clay, it could be subtracted mathematically from late-glacial pollen spertra. Vegetational history could then be studied as it was recorded by the primary pollen. Secondary microfossils in glacial drift have also been reported in this country. Mather, Goldthwait, and Thiesmeyn 11942) and Sayles and Knox (1943) concluded that microfossils in tills and glacial silty clays on Cape Cod, 25-50 miles east-southeast of Taunton, were deri \-ed from interstadial or in terglacial marine deposits that had been over-ridden and eroded by glacial ice. Recently most pollen investigators working with inorganic late- and postglacial sediments in America have considered the possibility that the sedi ments were contaminated with rebedded pollen. In most cases the criterion for suspecting contamination was the occurrence, in the same strata, of pollen grains from genera and species that are apparently ecologically incompatible * Now at Department of Botany, University of Michigan, Ann Arbor, Michigan. 211 212 Margaret B. Davis-The Problem of Rebedded Pollen in the present vegetation. In two instances (Andersen, 1954; Martin, 1958) pollen types which conflicted with vegetational and climatic interprt'tations based on the majority of the pollen in clays and sandy sediments, and which were therefore considered secondary, were subtracted from the pollen sum, and "corrected" pollen diagrams were published. In otht'r cases (Leopold, 1956; Livingstone and Livingston!', 1958; Da,·is, 1958; Ogden, 1959), pollen types for which a secondary origin was suspected were plotted on the diagrams but given less weight in the vegetational and climatic interpretation than other pollen types. In no instance, however, was the investigator able to demonstrate the source of the putative rebedded pollen, or to employ the objective subtrac tion method of Iversen ( 19:36). In the course of analyzing the pollen in a core from a bog ncar Taunton, l\Iassachusetts, the author found evidence of contamination of the late-glacial sediments with pollen rebedd~d from pre-Quaternary deposits. This paper re ports this evidence and presents the results of an investigation of the possible sources for rebedded pollen at Taunton. The problems of estimating the extent of contamination and of interpreting pollen analyses of contaminated sediments are discussed. ACK:'>OWLEDG'Il E'>TS I acknowledge with gratitude the many helpful suggestions and criticisms of Joseph H. Hartshorn and John P. Schafer. John C. Goodlett, Daniel A. Livingstone and Estella B. Leopold also JHovided valuable criticism of· an earlier version of the manuscript. The work \las done at the Biology Depart ment, Han·arrl university, and at the Geology Division, California Institute of Technology; it was supported by NSF fdlo\lships 470:)5 and 40002 and rt' search grant (to E. S. Deevey) G-4802. This paper is Contribution 979 of the Di,·ision of Geological Sciences, California Institute of Technology. E\~DE"CE OF REBEDDED POLLEN The pollt>n diagram from the Williams Street bog near Taunton, Massa chusetts (Davis, 1960) is similar in its major features to pollen diagrams from other parts of southern 1\ew England I Leopold, 1955; Leopold and Scott, 1958; Davis, 1958). It appears to reconl the dewlopment of forest during the Two CrePks interstadial (pollen zones T and A-1-2-:3), the decline in frequency of pine and deciduous trees relative to spruct> during the Valders ,stadia] (zone A-4), and the subsequent decline of spruce and increase in the frequt•ncy of pine and deciduous trees (zone B). (The upper levels of the deposit, in which deciduous tree pollen is presumably abundant. were not analyzed because of poor pollen preservation.) The stratigraphy of the core is as follows: 0-2.77 m Brown peat with wood and moss fragments. Loss on ignition 98-99% dry weight (2 samples). Pollen zone C (presumably). 2.77-2.86 m Coarse fibrous peat with grasslike leaf fragments. Loss on ignition 97% ( 1 sample I . Zone B-2. 2.86-:3.4-:3 m Very fine fibrous peat. Loss on ignition 52-8:'\;lo (2 sam ples). Zone B-2. :3.43-:3.62 m Very fine fibrous pt>at grading to gyttja. Loss on ignition :30% ( 1 sample) . Zone B-l. in Late-Glacial Sediments at Taunton, liJassachusetts 213 3.62-4.00 rn Gray-green gyttja. Loss on ignition 19-22% (5 samples). Zone A-4. 4.00-4.41 rn GrePn-brown gyttja. Loss on ignition 22-29% (5 samples). Zone A-1-2-:). 4.41-4.47 rn Fibrous pPat with grass-like lPaf fragments. Loss on igni tion 25'/o ( 1 sample I. ZonE' A-1-2-3. 4.47-4.95 m Gray silty sand streaked with dark gray sand. Sand at 4.80 m. Loss on ignition 2-4% (6 samples). Zone T?. 4.95-5.20 rn :\Tedium coarse sand. Lo~s on ignition l-2% (5 samples). Zone T?. It was not possible to penetrate deeper with the Hiller borer. A major difference between this dia!"ram and others in southern New England is the occurrence, in the lo\\·er levels of the deposit, of pollen of gen era that arc not found in the modern flora of the Vnited States. Pollen of Platycarya, a gPnus now limited to Asia. was found at 4.52, 4.85, 4.95, 5.05 (2 grains), 5.15 (7 !"rains), and 5.20 m 18 grains); Pterocarya, also limited to Asia. at 5.10 m 11 grain); and Engelhardtia-Alfaroa, limited 'to Asia and Central America, and to-Central America. respectively. at 4. . 80, 5.05 (:1 grains), 5.15 (2 grains) and 5.20 m 18 grains I. Plat_rcarya, Pterocarya, and Engel hardtia have all been reported from Tt·rtiary deposits in this country, and Pterocarya has also been reported from the Pleistocene of Maryland (LaMotte, 1952; Leopold and Scott ms.). As thesP genera do not occur in the modern flora, and are not n·corded in the post:rlacial fo~sil record, it seems improbable that they grew in the vicinity of Taunton at the time of the retreat of the ire. A far more plausible explanation is that their pollen has been rebedded from Tertiary deposits. If this is so, Tertiary pollen typPs that are indistinguishable from the pollen of modern speciE's may also he relwdded in the sediments in which these pollen types occur. The pn·,;ence of primary pollen as well is im plied by the general similarity of the pollen spectra at these levels to pollen zone T at other sites. To subtract secondary pollen from a miwd spectrum of primary and secondary pollen, it is necessary to know the complf'te spectrum either of the primary pollen or of the secondary pollen. As the purpose of the original in vestigation was to discover the primary pollen spectrum, no assumptions as to its nature are justified. It may be possible, however, to determine the secondary pollen spectrum from older pollen-hearin!" sediments in the vicinity, or from till or glacial clay deposited hy a glacier that had noded such sediments. Be cause the secondary pollen may have hePn derived from several different deposits, analyses of older sediments in the vicinity may not include all the secondary pollen types, nor represent them in the proper proportions. In any event, older pollen-lwaring sediments I other than varved clay) are not known in the Taunton quadrangle (Emerson, 1917). Difficulties may also be en countered in attempts to reco\·er the secondary pollen spectrum from glacial deposits. Iversen (19:36) found identical pollen spectra in till and glacial clay in Denmark and concluded that all the pollen in both types of deposit was secondary. This implies that plants did not li,-e within pollen dispersal distance when the ice was active and when the van·ed clay was deposited. It cannot be 214 Margaret B. Davis-The Problem of Rebedded Pollen assumed that this was true in ::\Iassachusetts, as late-glacial pollen diagrams record a comparatively rapid immigration of trees onto the newly deglaciated landscape. Thus supraglac:ial till and glacial clay in this region may contain primary as well as secondary pollen.1 In an attempt to determine the spectrum of secondary pollen rebedded in the Williams Street bog sediments, samples of nearby ground moraine, flowtill and varved clay were analyzed for their pollen content. POLLE:\" A:\"A.LYSIS OF TILL Four samples of unw(•athPred till were collected from borrow pits and building excavations in the vicinity of the Williams Street bog. The locations of the collections are shown in figure l. Samples A and B are flowtill, ablation moraine which slid as a mudflow from the surface of glacial ice onto stratified sand and gravel (Hartshorn, 1958). Samples C and D are ground moraine and contain abundant stones of the local bedrock." Fig;. l. Outline map of a portion of the Taunton quadrangle, showing locations at which samples were collected. Contour inlPnal 50 ft, datum mean sealevel. Inset shows approximate location in state of "'fassachusetts. Dra.m from U. S. Geological Survey topo graphic map, l :21000, rc\-iscd l9-l9. The samples were treated in the laboratory with HF and HCl, as recom mended by Heinonen ( 1957). Bromoform flotation (Frey, 1951) was also u~ed as a final step to remm-e the precipitate which formed in hydrofluoric acid. No pollen grains were found in samples B and D, although 1 grain (Fraxinus-type) was found in an empty tube put through the extraction pro cedure as a control for contamination at the same time that sample B was prepared. Three grains were found in sample C: 1 Fraxinus-type, 1 Ambrosia, and 1 Pinus. Unfortunately no control was run on the day this sample was 1 The occurrence of leaf and seed fossils of arctic-alpine plants in varved clay in the Connecticut Valley, :'\iassachusetts (Emerson, 1898) further ,indicates that varved clay may contain primary pollen. • I am grateful to J. P. Schafer for sending me till samples B, C and D. in Late-Glacial Sediments at Taunton, Massachusetts 215 prepared, but the occurrence of pollen in a predous control suggests that the sample was originally devoid of pollen. By contrast, Heinonen ( 1957), using much the same laboratory method and the same size samples, was able to extract from 0-9,340 tree pollen, 0-3,340 herb pollen, and 0-980 spores from samples of basal till from Finland. A few pollen and spores were extracted from sample A by means of repeated bromoform flotation of fines washed from a 600 gm sample. Heinonen (1957) has shown that this method is much less effecti,~e than HF for extraction of pollen from till, because most of the pollen occurs in a crust of fine particlE's adhering to the larger stones. A further dis advantage to the method is the increased possibility of contamination, as large amounts of reagents are used and the large sample is exposed to air for the several hours required for the procedure. Fifteen pollen and spores were found in sample A: 3 Ambrosia, 3 Betula, 1 Composilae, 1 Fraxinus, 1 Pinus, 1 un known, 1 Lycopodium, and 4 fern spores. Lnfortunately no control was run at the time the sample was prepared, but when a control was put through the same procedure some months later, 6 pollen grains were found. It seems likely that most of the pollen grains and spores found in this sample also represent laboratory contamination. E,~en if it is assumed that the till samples contain pollen in addition to laboratory contamination, the numbers of grains are so low that it seems un likely that either the till or the glacial ice from which it originated could ha,~e sen~ed as a principal source of secondary 'pollen. In any event, pollen of Platycarya, Ptcrocarra, and Engelhardtia-Alfaroa would be unexpected in the till, because the nearest known Tertiary sediments from which the pollen might have been derived are located approximately 25 miles to the northeast (Emer son, 1917). Tertiary pollen would occur in the ground moraine at Taunton only if ice had eroded these sediments an<l subsequently flowed 25 miles to the southwest. The general direction of glacial flow was apparently not toward Taunton, howe\'er, but to the southeast (Flint and others, 1959). POLLEN ANALYSIS OF Y~""RYED CL"" Y In contrast to the till, varved clay from the Taunton quadrangle contains abundant pollen. Samples were collected from the south waH of a claypit just east of the Taunton River (fig. 1). Sample E was collected about 8 ft below the surface of the day, 3 ft below an unconformity between gray varved clay and overlying brown oxidized clay, sample F from a thick varve 1 ft lower in the section, and sample G about 50 ft further west, 2 ft below a thick varve believed to be the same as that from which sample F was collected. Pollen was extracted without difficulty by flotation in bromoform, boiling in 10% KOH, and acetolysis, and was stained and mounted in glycerin. About 350 pollen and spores were counted from each sample. It is assumed that laboratory contamination was not significant; if any occurred it would represent only a small proportion of the abundant pollen extracted from each sample. Over fifty different pollen types were found in the varved clay. The per centages are shown in table l. (For comparison, the table also shows the pollen spectrum .from 5.20 m depth in the bog sediments.) The percentages of pollen in the three clay samples appear similar, but not identical. Differences in per- 216 Margaret B. Davis-The Problem of Rebedded Pollen TABLE 1 Pollen percentages in YarYed day samples, and sample from 5.20 m level at Williams Street bog Pollen or spore type Percentage in sample VarYed clay Bog sediment E F G 5.20m Abies ...................................................... 1.9 0.3 0.3 0.1 Acer saccharum-type .......................... .. 0.3 Alnus ...................................................... 4.0 1.4 :2.6 4.7 Ambrosia-type ....................................... . o.:-l 0.4 Anacardiaccae ...................................... 0 ..1 0 . .5 0.6 Artemisia-type ........................................ 0.5 u 0.5 0.3 Betula ...................................................... 11.1 23.1 27.n 6.1 Carya ...................................................... :i.:-l :.'.:2 1.3 Caryophyllaceae ................................... . 0.7 Castanea ................. ;.............................. .3.5 2.9 2.0 Cephalanthus-typc ................................ 1.3 0.6 0 . .5 0.3 Chenopodiaccae ................................... . 0 .. 5 0.1 Corylus .................................................... 4.0 1.7 :.'.9 1.5 Cupressincae .......................................... :.'.+ 1.4 0.5 0.2 Cypcraceae ............................................ 0.5 2.3 0.5 19.7 Engelhardtia .......................................... 1.3 1.0 Ericaceae ............................................... . l.l 0.5 2.0 Fagus ..................................................... . 0.3 0 .. 5 Fraxinus .................................................. 2.7 1.7 u u Gramineae .............................................. :2.9 2.5 1.3 6.9 /lex .......................................................... 0.3 0.3 0.3 Leguminosae'( ..................................... . 0.3 0.3 Liquidambar ......................................... . 0 . .3 Myrica .................................................... 0.3 0.6 0.3 0.3 Nymphaea .............................................. O.• ' l Nyssa ...................................................... :i.S 5.0 4.2 0.8 Ostrya-Carpinus .................................... 1.1 0.3 0.5 0.3 Picea ...................................................... 0 ..) 1.9 1.0 3.3 Pinus ...................................................... 8.5 8.1 1:.'.2 29.2 Plantago ................................................ . 0.1 Platycarya .............................................. 2.4 0.3 0.3 0.9 Platycarya or Engelhardtia ................ O.Ci Pterocarya .............................................. 0.3 Populus .................................................. 0..5 0.3 1.0 Potentilla ............................................... . 0.1 Quercus .................................................. n.6 1:2 . .5 16.7 1.3 Rosaceae ............................................... . 0.3 0.3 0.1 Rumex .................................................. .. 0.1 Sagittaria-typc ...................................... 0.:3 Salix ....................................................... . :2 .. 5 0.8 3.9 SymJJlocos ............................................. . 0.6 0.3 Thalictrum ........................................... . 0.9 Tsuga ...................................................... 03 0.6 0.:-l Tulmlillorae ............................................ 1.3 1.3 1.2 Typhaceac ............................................. . 0.6 0.3 Ulmus ..................................................... . 0.2 Umbellifcrae ......................................... . 0.6 1.0 0.3 Viburnum .............................................. 0 ..3 Vitaccae .................................................. 2.+ ,).3 0.3 Unknown .pollen .................................... 20.2 9.7 8.4 3.7 Unknown, type I .................................. 0.5 0.6 0.1 Filiccs ...................................................... 3. 7 6.7 2.6 1.0 lsoetes .................................................... .. 0.3 Lycopodium ........................................... . 0.3 0.1 in Late-Glacial Sediments at Taunton, M~assachusetts 217 TABLE 1 (Continuf'd) Pollen or- s~p-or-e -t~yp-e- -~--~~~~~~~- Percentage Ill sample Varved clay Bog Sediment E F G 5.20m L. selago-[ucidulum ............................. . 0.2 Sphagnum ............................................. . 1.6 3.3 1.3 1.1 .. :::::::::::::::::::::::::::::::::::: 1.3 0.6 0.1 YH;~:~;:~ ~~:~~~-~ 0.5 100% 100% 100% 100% Number pollen and spores counted .... 3/6 353 384 894 centages may be due to mistakes in identification, as well as to statistical error, as many of the pollen types found were unfamiliar to the author. Lower percentages of Betula and Quercus in sample E, however, may represent a real difference between this sample and the others. Of particular interest is the occurrence in the varved clay of Platycarya, Pterocarya and Engelhardtia-Alfaroa pollen. Photomicrographs of these types are shown in plate 1, figs. 1-6, 9. \Iany of the pollen types in the clay are un like those I have seen in Quaternary deposits in Massachusetts and are only tentatively identified or ,are listed as "unknown". Many of these unknown types were also observed in the hog sediments. For example, "Unknown, type I" (pl. 1, fig. 7) was found at 4.85, 4.90, 5.05 (2 grains), 5.10 ( 4 grains), 5.15 (2 grains), and 5.20 m. This pollen is characterized by very heavy exine thickenings around the pores ( colpae?) and appears simiiar to a pollen type found in upper Cretaceous and Eocene deposits and, as rebedded pollen, in postglacial marine cia ys in Sweden (Fries and Ross, 1950: cf. pl. 1, fig. 7). A similar pollen type, "Unknown, type II" (pl. 1, fig. 8), appears identical to pollen known from upper Cretaceous and Paleocene sediments in Germany, and, as a rebedded fossil, in late- and postglacial marine days in Sweden (Fries and Ross, 1950: cf. pl. 1, fig. 6'1. Several types of spherical microfossils with forked spines were found; one of these "Hystrix" types is shown in plate 1, figure 10. Similar microfossils are known from Cretaceous and Tertiary marine sediments in Europe, and have been reported in Quaternary tills and clays, where they apparently have been secondarily deposited (Iversen, 1936; Fries and Ross, 1950). "Hystrix" is thought to be a marine protozoan, but its exact affinities are unknown. It represents an ex-tant form, as appears to be in dicated hy its occurrence in postglacial freshwater peat, ncar a contact with salt-marsh peat, at the ] ohn Hancock fish weir site at Boston, :\Iassachusetts (Wilson, 1949). Apparently some, and perhaps all, of the pollen in the varved clay has been rebedded from older deposits. As the ice in the immediate vicinity seems to have contained little if any pollen, the secondary pollen must have been carried from some distance, presumably by glacial meltwater. This conclusion is supported by studit>s of the stones in outwash deposits at Taunton, which indicate meltwater transport from the northeast (J. H. Hartshorn, personal communication). In addition to pollen derived from Tertiary (and Creta- 218 Margaret B. Davis-The Problem of Rebedded Pollen 2 3 [I 0 fA- in Late-Glacial Sediments at Taunton, Massachusetts 219 ceous'?) sediments, the meltwater mav have transported pollen derived from interglacial and interstadial depo~its that were eroded by the ice, and pollen from vegetation then growing some distance from Taunton. It is also possible that the varved day contains primary pollen from plants growing within pollen dispersal distance of Taunton at the time the van·ed clays were deposited. A previous investigation of the vaned clays at Taunton (Hyyppa, 1955) led to quite different conclusions. Hyyppa identified pollen of twelve native tree genera in the van·ed day. He assumed all the pollen was primary and considered the predominance of deciduous tree pollen evidence of deposition during an interstadial when a "southern deciduous climate" prevailed not far outside the margin of the glacier ice. On the basis of the occurrence of marine, brackish water and a few freshwater diatom species in the varved clay he concluded that the clay was deposited in a shallow marginal sea. However, I consider the presence of pollen of exotic genera, even though it occurs in low percentages, evidence that the clay is contaminated with rebedded microfossils. For this reason the microfossils are unreliable as vegetational and climatic in dicators of the conditions that prevailed at the time the clay was deposited. DISCl'SSIO" The sediments in the Williams Street bog apparently were contaminated with secondary pollen either through the erosion and redeposition of varved clay, or through the direct action of meltwater streams flowing into or over the site of deposition. The first of these hypotheses seems unlikely because the surface of the varved clay is at least 10 to 20 ft lower than the esker and kame or lake deposits that surround the bog. However, it is conceivable that other outwash deposits (e.g. the kame deposits?) were eroded and redeposited as the ice block underlying the site of deposition melted. If these outwash deposits contained pollen from meltwater streams deposited contemporaneously with the poHen in the varved clay, the spectrum found in the varved day would be similar to the spectrum of pollen deposited secondarily in the hog sediments. Even when these assumptions are made, however, it is still not possible to ap· PLATE l l\Iicrofossils from Yarwd clay sample E. The photomicrographs were all made with an oil immersion lens, and ha\e been equally enlarged. Scale indicated in lower right hand corner of plate. Figs. l-3, Platycarya. Fig. I, focus on upper surface of grain, Fig. 2, focus on optical cross section of pores, Fig. 3, focus on lower surface of grain. Figs. 4-5, Platycarya-type ? Fig. 4, focus on optical cross section of pores, Fig. 5, focus on lower surface of grain. Fig. 6, Engelhardtia-Alfaroa ? , focus on wall, and optical cross section of pore in upper portion of photograph. This grain rPsercbles d. Engelhardtia pollen from the late Tertiary of Denmark (Ingwersen, I95S). Other grains seen resembled modern pollen more closely. Fig. 7, "Unknown, type I", focus on surface of wall in upper left-hand portion of grain, showing pore (col pus?), and on optical cross section of upper right-hand pore, showing heavy cexine thickenings. Fig. R, "Unknown, type II"', focus on optical cross section of colpae, showing heavy exine thickenings. Fig. 9, Pterocarya, focus on uppermost, folded portion of grain. Only one of the fiye pores is clearly in focus. Fig. 10, "Hystrix", focus on two of the many forked spines. Other "Hystrix" types seen were larger, with heavier spines. 220 Margaret B. Davis-The Problem of Rebedded Pollen ply h-ersen's (1936) subtraction method. hersen divided the secondary pollen spectrum into two groups: Group A included poHen of extinct Tertiary genera, and Group B included pollen which could not be distinguished from Quater nary pollen types. He demonstrated that the ratios of pollen types within Group A remained constant in late-glacial sediments, regardless of the total amount of Group A pollen present. Assuming, therefore, that the ratio of Group B pollen to Group A pollen had also remained constant, he subtracted pollen types belonging to Croup B in the ratio to Croup A in which they had occurred in the till. Group A pollen was also subtracted, and the remaining pollen was assumed to be of primary origin. This method cannot be used here because the pollen of exotic genera (Group A i occurs in low frequencies both at the bog and in the varved clay deposit; the ratio of Group B to Group A pollen cannot be calculated with statistical accuracy. The sedimPnts may also have been contaminated by a meltwater stream flowing over the site before the underlying ice had melted, or into the site after melting had begun. J n this east> the spPctrum of pollen in the varvt>d clay is not necessarily the !'arne as the secondary pollen spectrum in the bog sediments, as the van-ed clay may contain both primary pollen, and secondary pollen brought in by other meltwater streams. Assuming the hog sediments were contaminated directly by glacial melt\rater, they may be of the same age or only slightly younger than the UJlJWrmo,-t levels of the varved clay deposit. If samples had been collected hight>r in the day section, it might have been possible, by comparing the pollen spectra \\ith those found in the lowest levels of the hog deposit, to test this hypothesis. Direct contamination by glacial mPitwater is in contrast to the mode of contamination described in Europe. In Denmark and Finland the boulder-clay is often rich in pollen (ht>rsen, 1936; Heinonen, 1957). Pollen and fine in organic debris wPre apparently washt>d from the moraines and redeposited in lakes and ponds during pt>riods of inten,:e solifluction. Autochthonous, organic late-glacial sediments, depositPd when slope-wash was not effecti\-e, -are oftf'n entirely free of secondary pollen; usualh- the amount of contamination varies with the clay content of the sediments I hersen. 19:16. 1947; Krog, 1954). At Taunton, however, the erosion and redeposition of till and outwash deposited by local ice would not have resultt'd in extensive deposition of secondary pol len. If the deposition of secondary pollen depended upon meltwatPr streams. there is no necessary correlation between the amount deposited and the loss on ignition of the sediments, nor between the amount deposited and the seYerity of the climate. As the spectrum of secondary pollen deposited in the bog sediments has not been determined with certainty, the spectrum of primary pollen remains unknown. "Mixed primary and secondary pollen spectra are meaningless-they rt>present vegetation that never existed at a point in time, and vegetational interpretations cannot be attempted. Above the 4.52-m level, however, pollen of Platycar;·a, Pterocarya, Engelhardtia-.4lfaroa and "Unknown, type I" were not found. The sediments above 4.52 m may be free, or nearly free, of con tamination, and a vegetational interpretation based on the major constituents ran be made with reasonable confidenre.