eevee damellmim 6 nm latitude in the cold Labrador water found between the Gulf Stream and the coast of North America. In the mid-Wisconsin it ranged as far south as 13°N latitude along the African coast (Fig. 1). The average latitude shift between recent and glacial is approximately 15° latitude and even greater between living material and glacial (4). At the opposite extreme in temperature tolerance is Umbellosphaera irregularis (Fig. 1), a tropical species. The minimum shift between recent and glacial boundaries for this form is approximately 20° latitude. The accuracy to which such shifts can be determined reflects the spacing of core samples. Between the glacial presence (core Nos. V16-200 and A180-72) and absence (core No. A180-56) there is a gap of 10° latitude (see Table 1). The temperature ranges of Helico- ee me ee water forms. The distribution of U. mirabilis is today bounded by the 18°C isotherm; this may be too high a value since the colder-stage coccoliths are present at higher latitudes in sediment samples from this study than was ob- with the higher percentage (which today occurs in this species in transitional waters) indicative of the optimum range. served with either living or surface sedi- consin. This type of subtle biogeograph- ment material in the survey of modern forms. It is possible that the colder stage of U. mirabilis has a distribution similar to that of R. stylifera and S. pulchra. Nevertheless, the species shows a definite shift between recent and glacial sediments which averages 15° latitude (Fig. 2). Further evidence indicative of a cooling of the Atlantic in glacial times is the preponderance of cold-form coccoliths in glacial-age sediments from the mid-North Atlantic, a situation reversed in recent surface sediments. Cyclococcolithus leptoporus (Fig. 2) is among the most eurythermal of the 1). They are presently Coccolithophoridae. Today it ranges from the equator to Arctic waters. There is no apparent difference in maximum range between recent and distribution during mid-Wisconsin time. to the lack of core material further north than the line Sp 10-1 through sphaera carteri, Rhabdosphaera stylifera, and Syracosphaera pulchra are not as limited as those of the two preceding species (Fig. found in subtropical to transitional waters. All show a much more limited Helicosphaera carteri has an average distributional difference of 10° latitude while Rhabdosphaera stylifera and Syracosphaera pulchra both have an glacial, although this is probably due R 10-2 (see Table 1). There is a marked change in percentage distribution between recent and glacial populations, approximate shift of 17° latitude. Note that in these three species the recent maximum distributional lines follow the northern border of the subtropical gyral (Gulf Stream) across the Atlantic Ocean. This agrees with plankton data from the North Atlantic, where the boundary between subtropical and subarctic species approximates the northern border of the Gulf Stream (4). Umbellosphaera tenuis and Umbili- cosphaera mirabilis, while having a sat- isfactory preservation record, do not show as distinctive a difference in their recent and glacial distributions as the preceding species. Umbellosphaera tenuis is found at higher latitudes and is limited today by the 16°C isotherm; it is typical of subtropical waters. Its glacial to recent shift is from 10° to 15° latitude (Fig. 2). It is not as abundant in glacial as in recent sediments. This may be due to the fragility of its macrococcoliths. The coccoliths of Umbilicosphaera mirabilis, like Coccolithus huxleyi, have temperature-dependent structural variations (4, 6). While these changes are gra- dational in C. huxleyi, U. mirabilis appears to have separate cold- and warm1316 Table 1. Core locations in the North Atlantic and the depth of the glacial sample in each core. Depth Core No , Location Latitude Longitude of glacial sample (cm) A153-146 Al56-4 Al156-5 Al64-59 A167-13 A167-14 A179-13 A1T79-17 AT80-9 A180-16 A180-32 A180-48 33°43'N 34°49 'N 37°07'N 38°42'N 31°39'N 31°28'N 23°56'N 28°00'N 39°92TN 38°21N 29°O7'N 15°19’'N 44°45'W T4°41'W 73°37W 67°52'W 75°21'W 76°28'W 78°45W 73°47 W 45°STW 32°29W 26°15'W 18°06'W 80 844 95 245 300 300 97 280 115 140 59 488 A180-72 R5-36 R5-54 R5-57 R10-2 SP8-4 SP9-3 SP9-4 SP10-1 V16-200 00°36'N 46°55'N 25°52'’N 19°40'N 56°59'N 32°50'N 53°53'N §0°02'’N 51°23'N 01°58'N 21°47'W 18°35'W 19°03'W 19°06'W 12°238'W 18°32'W 21°06 W 14°46'W 38°04"W 37°04 W 120 162 35 270 100 65 220 200 150 120 A180-56 12°15'N 17°46'W 207 Thus the line representing percentage change may be a rough indicator of the subtropical boundary in the Wis- ic change will require further work before definite conclusions can be drawn. A number of other species showed some degree of biogeographic change by a shift in maximum boundaries: however, they are not plotted because of relative rarity in the core material. In the case of the ubiquitous eurythermal species (Coccolithus huxleyi and Gephyrocapsa oceanica), their distri- bution is similar to that of Cyclococcolithus leptoporus. One interesting change, presently not usable for paleoclimatic work but im- portant in systematics, is the reversal in dominance of Coccolithus huxleyi and Gephyrocapsa oceanica from glacial to recent (Table 2). In today’s ocean Coccolithus huxleyi usually constitutes over 50 percent of the flora, but in the mid-Wisconsin it shared and in some latitudes was dominated by Gephyrocapsa oceanica. I consider that Coccolithus huxleyi is a relatively recent form, none being found before the Pleistocene, and that it evolved from the Gephyrocapsa oceanica complex during the late Pleis- tocene. Similarities in form and ecology. combined with the finding of intermedi- ate forms in Pleistocene core samples that T am now investigating, support this theory. A comparison of the latitudinal change in flora from northern cold wa- ters to southern warm waters in recent and glacial sediments indicates the dominance of cooler water forms in lower latitudes during the mid-Wiscon- sin. Umbilicosphaera mirabilis and Syra- cosphaera pulchra constitute a much larger percentage of the flora in glacial times in this area, while Cyclococcolithus fragilis and Umbellosphaera ir- regularis, subtropical to tropical forms. are nearly absent during glacial time from the North Atlantic. The biogeographic boundaries of Coc- colithophoridae species in today’s seas can be correlated with surface water isotherms (4). If we assume that the present temperature ranges of the spe- cies held for the last glacial period. then the paleogeographic boundaries of species can also be assigned temperature values. If all these boundaries are SCIENCE, VOL. 158