SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY NUMBER 4 Richard Gfeiti Distribution of and Roberta K. Smith -p^ -, • i-» • r rlanktonic r oraminilera in the Vicinity of the North Atlantic Current ISSUED APR 13 «70 SMITHSONIAN INSTITUTION PRESS CITY OF WASHINGTON 1970 ABSTRACT Cifelli, Richard, and Roberta K. Smith. Distribution of Planktonic Foraminifera in the Vicinity of the North Atlantic Current. Smithsonian Contributions to Paleo- biology, 4:1-52. 1970.—Planktonic Foraminifera collected from the vicinity of the North Atlantic Current and the Gulf Stream during late winter-early spring and fall of 1964 are described and their distributions are recorded. Variations in faunal composition seem to be related largely to water regime dynamics and seasonal cycle. Among the fall collections, three distinctive assemblages can be recognized : a western group in the vicinity of the Gulf Stream, containing predominantly Sargasso Sea-Gulf Stream species dominated by Globigerinoides ruber; a northern group, dominated by Globigerina quinqueloba egelida, new subspecies, reflecting the influence of cold, northern waters adjacent to the North Atlantic Current; and an eastern group, dominated by Globigerina incompta, apparently developed within the limits of the North Atlantic Current. The last group seemingly represents an anomaly, as North Atlantic Current surface temperatures were relatively high at the time of collection, and dominance of a warm-water form, such as Globigerinoides ruber, might have been expected. The anomaly suggests that the North Atlantic Current is a partially closed gyre, fed by both slope waters and Gulf Stream. Temperatures are considered to be close to threshold for both cold and warm-water species. Distributional patterns displayed by the late winter-early spring collections are compatible with the proposed model. Also, these collections, taken over a period of almost three months, reflect marked seasonal changes in faunal composition, par- ticularly in Sargasso Sea-Gulf Stream elements. Twenty-five species and subspecies are described. One species, Globigerina atlantisae, and one subspecies, Globigerina quinqueloba egelida, are new. Official publication date is handstamped in a limited number of initial copies and is recorded in the Institution's annual report, Smithsonian Year. UNITED STATES GOVERNMENT PRINTING OFFICE WASHINGTON : 1070 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price 65 cents (paper cover) Contents Page Introduction.................................. 1 Acknowledgments................................ 1 Methodology.................................. 2 Hydrography.................................. 2 Distributional Data............................... 5 Atlantis 77-13 Distributional Patterns................. 6 Numerical Abundances and Diversity.................... 6 Stations 2-13, 18, 19, 21 .......................... 9 Stations 16, 26, 28, 29........................... 9 Stations 32-42............................... 10 Summary of Atlantis 77-13 Distributional Patterns and North Atlantic Circulation................... 11 Western Stations............................. 11 Eastern Stations ............................. 11 Dynamics of Plankton Populations..................... 12 A Distributional Model for the North Atlantic Current.......... 12 Atlantis 77-9 Distributional Patterns................. 13 Numerical Abundances and Diversity.................... 13 Stations 286, 288.............................. 13 Stations 327, 337, 345, 347......................... 13 Stations 385-408.............................. 13 Measurement of Chambers and Test Volution.................... 15 Systematic Descriptions............................. 17 Literature Cited................................ 43 Plates 1-6................................... 45 Index .................................... 51 m Richard a/ein Distribution of and Roberta K. smith planktonic Foraminifera in the Vicinity of the North Atlantic Current Introduction The North Atlantic Current is that part of the North Atlantic gyre formed south and east of the Grand Banks. Although the North Atlantic Current represents, in part at least, a northern continuation of the Gulf Stream, it manifests a distinct hydro- graphic setting. In this paper, planktonic Foraminifera from the vicinity of this current system are described and their distributions recorded. In addition, we have attempted to discern distributional patterns and ex- plain them. This study is based on plankton collections obtained from two cruises of the Woods Hole Océanographie vessel R/V Atlantis II in 1964. The first cruise, At- lantis II-9, occupied plankton stations during winter- early spring between 1 February and 29 April in the region generally south and east of the Grand Banks. The second cruise, Atlantis 77-13, occupied stations in the same general region in fall, between 2 and 21 Sep- tember. In addition, stations were occupied during Atlantis 77-13 west of the Grand Banks, along the Gulf Stream's mean path. Figures 1 and 2 show sta- tion locations from both cruises with respect to the major circulatory features of the North Atlantic. As our studies of North Atlantic planktonic Forami- nifera progress, we become increasingly impressed with the complex dynamics of the distribution of planktonic organisms. Owing to the environment's Richard Cifelli, Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Wash- ington, D.C. 20560. Roberta K. Smith, Department of Paleobiology, National Museum of Natural History Smith- sonian Institution, Washington, D.C. 20560. mobility, planktonic organisms are constant, involun- tary travelers (if we may be permitted an anthropo- morphic metaphor) that, during their lifetimes, may find themselves in places they do not care to be. There- fore, it is difficult to attach spatial limits to species dis- tributions or to relate these directly to specific-physio- chemical factors, such as temperature. Attempts have been made to do this (e.g., Bé, 1968; Boltovskoy, 1968), but these schemes, in our view, both oversim- plify the realities of nature and fail to distinguish among spatial, temporal, and physiochemical aspects of the environment. Information is still insufficient to attempt a general synthesis of planktonic foraminiferal distribution in the North Atlantic. Therefore, we have, in this study, limited our interpretations to particular distributional situations. Because the two cruises offer distinctly dif- ferent distributional data, we describe Atlantis 77-9 and Atlantis 77-13 separately. We develop our dis- tributional concepts along with the descriptions. Acknowledgments The plankton samples described were collected on cruises 9 and 13 of R/V Atlantis II of the Woods Hole Océanographie Institution; it is a pleasure to thank the scientific parties, officers, and crews during these cruises for their assistance. Financial support of the work at sea was variously by the Office of Naval Re- search (contract Nonr-2196(00) ), the Atomic Energy Commission (contract AT (30-1)-2174), and the Na- tional Science Foundation (grant GP 861) at the Woods Hole Océanographie Institution; processing and distribution of the samples were supported by 1 A.E.C. under the contract named. We thank each of these agencies for its consideration. V. T. Bowen of the Woods Hole Océanographie Institution has read the manuscript, as have F. L. Parker and W. Berger of the Scripps Institution of Oceanography; we thank them. Rudolpf S. Sheltema of the Woods Hole Océanographie Institution pro- vided useful information on circulation patterns of the North Atlantic and distributional patterns of pelagic larvae. Brenda Williams assisted in preparation of samples. Marsha Jessup drew the plates of Foraminif- era and L. B. Isham prepared the figures. To these and those other people who have given assistance, we are most grateful. This is contribution number 2370 of the Woods Hole Océanographie Institution. Methodology Samples mainly were obtained in oblique tows from between 200 and 300 meters with a number 10 plank- ton net (0.158 mm aperture) having a 94-meter open- mouth diameter. The ship's towing speed was between one and two knots. A few samples were obtained from between zero and 5 meters depth while the ship drifted on station. Figure 4 shows location, time, and depth of collections. Samples were preserved in 5 percent for- malin, buffered with hexamethylenamine. After arrival at the laboratory, samples were pre- pared and concentrated by the ignition method (Sachs, Cifelli, and Bowen, 1964; Sachs, 1965; Smith, 1967). Briefly, (1) formalin is washed out of samples, (2) samples are digested in sodium hypochlorite or hydrogen peroxide, (3) the digesting agent is washed out and the samples dried, (4) samples are ignited at 500 °C in a muffle furnace for approximately 1 to 2 hours, (5) the ashy residue is washed out of ignited samples, and (6) dried samples are stored in pyrex petri dishes for subsequent study. A random specimen count is needed, both to deter- mine specimen number per given volume of sea water and to be sure a taxonomically and numerically rep- resentative sample is seen and/or picked. Therefore, we follow a standard procedure. First, if the prepared sample is so large that it overcrowds the picking dish— either making specimens more than one layer thick or so dense that it is difficult to count specimens in a given area of the dish—we split the sample to an ap- propriate size with a microsplitter. SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY We use a rectangular metal picking dish with a grid of 100 divisions (Smith, 1967). If the sample contains less than 500 specimens, all are examined, counted, picked, and identified. If more are present, we select a random sample by using a table of random numbers to indicate the particular numbered rectangles in the picking dish in which specimens are to be counted. Thus, by obtaining a continuing mean as specimens from different randomly selected rectangles are counted, examination of 10 to 20 rectangles usually suffices to determine the total specimen number in the sample or per given volume of sea water. We usually pick and identify between 300 and 400 speci- mens (every specimen in the randomly designated rec- tangles ). Occasionally, it is necessary to pick more. Care should be exercised to spread the sample as evenly as possible in the picking dish so as to reduce the num- ber of rectangles necessary for consideration to reach a nearly constant mean. Subsequent to picking, all specimens are arranged according to species in an assemblage slide. Numbers of individuals of each species or taxon are then counted. From these numbers, absolute and relative abundances are calculated by comparing with the total number in the slide and sample. Hydrography In Figures 1 and 2 the Atlantis 77-13 and -9 stations are plotted on Sheltema's (1968, unpublished data, W.H.O.I.) North Atlantic circulation scheme. This compilation appears a reasonable compromise of pre- viously proposed circulatory systems (Stommel, 1958; Worthington, 1962; Mann, 1967). It shows North Atlantic circulation as essentially a single, clockwise, asymmetrical gyre, with the Gulf Stream swinging southeast at about longitude 50° W and the North Atlantic Current forming a more or less separate eddy of restricted dimensions. Stommel (1965) gives a thorough account of the western North Atlantic hydrography, particularly re- garding boundary conditions. The principal circula- tory feature is the Gulf Stream, which originates in the Straits of Florida and flows clockwise and north- easterly between the Sargasso Sea east and south, and the slope waters west and north. It does not move in a straight course, but flows in meanders which some- times develop into detached eddies, and the position varies appreciably throughout the year. The Gulf NUMBER 4 Figure 1.—Atlantis 11—13 stations plotted with respect to surface circulation (after Scheltema, unpublished data, Woods Hole Océanographie Institution). Stream, therefore, is a dynamic but effective boundary between the distinctive cold slope waters and warm Sargasso Sea. It shows a strong temperature gradient at the surface and in subsurface. Stommel (1958, p. 173) states, "The Gulf Stream is not a river of hot water flowing through the ocean, but a narrow ribbon of high-velocity water acting as a boundary that pre- vents the warm water on the Sargasso Sea (right-hand) side from overflowing the colder, denser waters on the inshore (left-hand) side." The concept of the Gulf Stream as a boundary rather than a river of water has important bearing on interpretation of distribution of planktonic organisms in the North Atlantic. The Gulf Stream can be traced about as far as longi- tude 50 °W where it passes southeast of the Grand Banks. From there on the continuation appears to con- sist of several distinct currents, but the nature of these currents still is obscure. Apparently, the Gulf Stream splits into two main branches ( Sverdrup, Johnson, and Fleming, 1942, fig. 187; Stommel, 1965). The stronger branch is diverted north as the North Atlantic Current. The southern branch moves diff usedly clockwise south- east and blends with the Canaries Current east of the Azores. According to this view, the North Atlantic comprises a single gyre with no sharp eastern boundary of the Sargasso Sea. Worthington (1962) proposes a different scheme for eastern North Atlantic circulation, whereby the Gulf Stream turns entirely southeast after passing the Grand Banks, completing a western Atlantic gyre. Northeast of this gyre is another, separated from the first by a low pressure trough over the southeastern SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Figure 2.—Atlantis II—9 stations plotted with respect to surface circulation (after Scheltema, unpublished data, Woods Hole Océanographie Institution). Newfoundland Ridge. This trough is considered per- manent, so that two separate gyres maintaining two more-or-less distinct water regimes compose North Atlantic circulation. Mann (1967) proposes a scheme that, in some ways, is a compromise of the previous ones. Accordingly, the Gulf Stream does swing south at about latitude 50°W, as suggested by Worthington. Mann, however, rejects the North Atlantic Current as part of a separate gyre. He proposes, instead, that both slope water and a northern branch of the Gulf Stream, flowing north and east, feed the North Atlantic Current. It is true, how- ever, that Gulf Stream and northern-water sources of the North Atlantic Current are still poorly known. Mann also suggests that the Gulf Stream swings sharply south at about latitude 35 °N and longitude 40° W. Unfortunately, Mann's data do not extend east of longitude 40 °W, and thus do not pertain to the eastern Atlantis 77-13 and -9 stations. His scheme, however, does suggest that the North Atlantic Current forms a partially closed eddy northeast of the main gyre. As both Gulf Stream and slope waters feed the eddy, a water regime results with planktonic popula- tions displaying unique dynamic relationships, as dis- cussed later. The North Atlantic Current forms a devious path of the North Atlantic gyre that eventually blends with the Canaries Current to the east and the Sargasso Sea to the south, at about latitude 35 °N. Therefore, the essential continuity of North Atlantic circulation as a single clockwise gyre is maintained, and the North At- lantic Current behaves as an extension of the Gulf NUMBER 4 Stream as a transport medium. The continuity of Gulf Stream-North Atlantic Current circulation is evi- denced by drift bottles released from eastern North America and recovered in the Azores and the trans- oceanic occurrences of numerous pelagic larvae of tropical shallow-water benthonic invertebrates (Shel- tema, personal communication). Many pelagic larvae have been recovered along the general course of the North Atlantic Current. At the same time, however, the North Atlantic Current is distinct from the Gulf Stream. A significant feature of the Gulf Stream is that it borders the Sargasso Sea. No part of the North At- lantic Current eddy is comparable to the Sargasso Sea as the 18° water, a distinctive feature of the Sargasso Sea (Worthington, 1959) is lacking there (Sverdrup, Johnson, and Fleming, 1942, fig. 186). Distributional Data Before summarizing the Atlantis 11-13 and -9 data and the distributional patterns implied by them, we will review some previous North Atlantic distributional data. While these data are few and mostly limited to the Western North Atlantic, they serve as a useful framework for interpreting the Atlantis 11-13 and -9 faunas. A study of four seasonal traverses across the Gulf Stream (Cifelli, 1962) strongly suggests that the west- ern North Atlantic could be viewed as containing two distinct endemic faunas. One consists essentially of Globigerina species and is found in the slope waters north of the Gulf Stream. The other consists of a di- verse group of species belonging to several genera, but with numerically few representatives of Globigerina. This fauna occurs in the Sargasso Sea and Gulf Stream. The Gulf Stream also must contain elements carried from the Caribbean Sea, but thus far distinctly diag- nostic Caribbean elements have not been recognized. The Gulf Stream fauna conforms with the concept that the Gulf Stream forms the western border of the Sar- gasso Sea (Stommel, 1965). In slope waters adjacent to the Gulf Stream, Globig- erina species are found with Sargasso Sea and Gulf Stream species. This is the boundary fauna and gener- ally, the frequency of Globigerina species increases with distance north of the Gulf Stream's mean path. The boundary between the Gulf Stream and slope waters, however, cannot be defined rigidly because its extent and position vary considerably throughout the year. 359-866 O—70------2 Seasonal faunal boundary variations most likely result from changes in position and extent of the Gulf Stream. According to the concept of two endemic western North Atlantic faunas, assemblages in the slope-water boundary are regarded as faunal mixtures rather than faunal transitions. Although the two western North Atlantic faunas appear to maintain their identity throughout the year, significant seasonal variations in species frequency-relationships occur (Bé, 1960b; Cifelli, 1962, 1965). Although these variations are not fully understood, particularly those occurring north of the Gulf Stream, some known aspects bear on the in- terpretation of the Atlantis II traverses. In the Sargasso Sea-Gulf Stream fauna, a spectacu- lar change in species dominance occurs between sum- mer and winter. In summer and fall, Globigerinoides rub er is dominant, but after the fall turnover it strongly declines, along with other Globigerinoides species. In winter, G. ruber appears mostly in negligible percent- ages, while Globorotalia truncatulinoides is dominant. During late winter, G. truncatulinoides declines and G. hirsuta achieves dominance. In spring, Globigeri- noides ruber again becomes dominant, but Globigeri- nella aequilateralis also occurs in high frequencies. The extent of slope-water fauna seasonal change is much less clear than in the Sargasso Sea-Gulf Stream fauna. Partly, the obscurity results from the slope- water fauna's greater degree of local lateral variation, which, in turn, may result from the slope-water envi- ronment's greater variability than the Sargasso Sea. Yet, past work (Cifelli, 1965, pp. 5-8) reveals some spatial and temporal frequency changes of Globigerina species which might help explain some Atlantis 77-13 and -9 frequency relationships. In previous traverses across western slope waters, maximum concentrations of Globigerina inflata were found during fall in inner slope waters, relatively close to the Gulf Stream (Cifelli, 1965, p. 6). Bé and Hamlin (1967, p. 102) also found G. inflata in high concen- trations during summer in inner slope waters and con- sidered it an indicator of "transitional waters." They also found it considerably less common in the eastern Atlantic in summer. Globigerina quinqueloba egelida, new subspecies, exhibited peak development in western slope waters during winter (Cifelli, 1965, p. 7) and was relatively scarce during summer. Bé and Hamlin (1967) found G. quinqueloba egelida maxima during summer in the subarctic region. G. incompta has exhibited maxima SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY in western slope waters during summer and fall, with frequencies being generally higher farther north of the Gulf Stream than those of G. inflata. From the above data it is clear that the slope-water and Sargasso Sea-Gulf Stream faunas each undergo significant changes in species frequencies during the year. In comparing faunas, one should take into account these seasonal changes. The Atlantis 11-13 and -9 collections were made when North Atlantic waters were, respectively, warm- est and coldest. These collections provide some infor- mation on seasonal faunal changes in the vicinity of the North Atlantic Current which is needed for mean- ingful comparison of this part of the North Atlantic water regime with that farther west. Because of the marked difference in seasonal setting of the Atlantis 77-13 and -9 cruises, they will be treated separately in the sections that follow. The check list (Figure 4) gives data on the occurrence and numerical abundance of all species from both cruises. Atlantis 77-13 Distributional Patterns The Atlantis 77-13 traverse covers much of the breadth of the North Atlantic, transecting a region of complex circulation. Changes in faunal associations are complex and difficult of interpretation because of the complexities of the planktonic environment and circulation in the region of the traverse; however, the tendency toward mutual exclusivity of Globigerinoides ruber, Globigerina incompta, and G. quinqueloba egelida, new subspecies, reveals some pattern. Figure 5 shows this, where the frequency of each is shown as a percent of the total of the three, not the total foraminiferal populations. Three more-or-less distinct populations are suggested, with only stations 4 and 29 appearing transitional. Moreover, these station group- ings by frequency are spatially compatible (Figure 1 ). Therefore, stations are grouped spatially and evalu- ated in light of the three distributional patterns sug- gested by the frequency relationships. As the stations were occupied in September, the summer (Cifelli, 1965; Bé and Hamlin, 1967) and fall (Cifelli, 1965) seasonal data are used as references. Numerical Abundances and Diversity Numerical abundances vary from about 200 to 80,000 specimens per 1,000 cubic meters of sea water. Most values compare favorably with those found previously in western Atlantic slope waters (Cifelli and Sachs, 1966). The western part of the traverse (stations 2, 4, 5, 8), however, shows counts less than 6,000 per 1,000 cubic meters, compared to previous counts in Globigerinoides ruber Globigerina incompta Globigerina quinquelobaegelida Globigerina inflata Globigerina bulloides sensu loto Figure 3.—Numerical abundance relationships among Globigerinoides ruber, Globigerina incompta, G. quinqueloba egelida, new subspecies, G. inflata, and G. bulloides, sensu lato from Atlantis 77-13 stations. LAIN LONG.W 4I°33' 64°37 4I-34' 62°00' 41-30' 60°I3' 41-15' 5r30' 4I°23' 55"52' 4I°38' 54-44' 4I°30' 50-00' 4I°30' 47-30' 4I°29' 47-04' 42^9' 45°52' 45°29' 43°06' 47°I0' 41-42' 47°28,47"09' 40*46' 38-501 46°I7' 35°32' 45-58 34-53' 41-33' 29-43' 41-33' 29-43' 41-04' 29-12' 52-55' 26-45' 52-09' 27-51' 37-12' 46-45' 41-15' 42-00 45°24 36°58 46*13' 35-33' 39-11' 27-23 45°55 38-22' 45-52 41-00' 45°23' 43-33' 45-10' 45°39' SPECIES STATIONS 2 4 5 8 9 13 16 IB 19 21 26 ?R 29 3? 34 36 38 *---- 40 AÍI3 42 Afl5-?86 -----1» ?88 327 337 345 347 385 397 400 404 408 Globigerina atlantisae 0 0 1 1 0 1 Orbulina universa 2 0 Probably most significant is the frequency relation- ship between Globigerina bulloides, sensu lato, and G. inflata. At both stations, G. bulloides, sensu lato, strongly dominates, with low percentages of G. inflata. At Atlantis II-9 station 404 the relationship is reversed ; G. inflata strongly dominates (69 percent), with few G. bulloides (3 percent). Nearby station 400 shows nearly identical frequency relationships, although with some increase of G. bulloides, sensu lato. NUMBER 4 15 Frequency relationships within the North Atlantic Current are not homogeneous, however. Station 397 shows almost equal percentages of G. bulloides, sensu lato, and G. inflata, and no species exhibits strong dominance. Station 385, at the southern end of the North Atlantic Current, reveals only 9 percent G. bul- loides, sensu lato, and shows dominance of Globiger- inita glutinata (36 percent). This variation's signifi- cance is not clear. Curiously, stations 385-408 are impoverished in Sargasso Sea forms. Globigerinita glutinata, a ubiqui- tous form, is the only species not of Globigerina that oc- curs in frequencies of over 10 percent. The Sargasso Sea form Globorotalia hirsuta composes 6 percent of the station 404 assemblage. At all other stations, com- bined Sargasso Sea forms account for 3 or less percent. Since we have no quasisynoptic data from the Sargasso Sea, we can only speculate on the cause of this impoverishment. Conceivably, however, April is a critical time in North Altantic planktonic population dynamics. Previous April data indicate that Globorota- lia hirsuta frequency declines rather sharply and forms such as Globigerinoides ruber and Globigerinella aequilateralis achieve dominance; however, April temperatures in the North Atlantic Current are still relatively low. Perhaps Globigerinoides ruber and Globigerinella aequilateralis lack as low a temperature tolerance as Globorotalia hirsuta; the former species develop maximally in the northern Sargasso Sea be- tween spring and fall and are virtually absent there in winter. Accordingly, in the context of the dynamic model proposed for Atlantis 11-13 distribution (page 28), we suggest that temperatures are generally below threshold in the North Atlantic Current for Sargasso Sea forms and they mainly are unable to survive the journey around the eddy. Measurement of Chambers and Test Volution Among planktonic foraminiferal taxa, particularly of Globigerina species, differences often are subtle and occasional morphologic overlap occurs. Because of this, we add to some species descriptions data on test size and number of chambers relative to test volution. (The species are Globigerina atlantisae, new species, G. bul- loides bulloides, G. dutertrei, G. incompta, G. aff. G. pachyderma, and G. quinqueloba egelida, new sub- species.) Such data point up resemblances and differ- ences among species and assist in establishing popula- tion maturity and geographic variation. We hope these data will allow comparisons in future studies. In descriptions of coiled Foraminifera, one custo- marily includes maximum diameter, total number of chambers (if determinable) and number of chambers occupying the periphery. Only rarely is the relationship between(l) size and/or number of chambers and (2) coil or degrees of volution examined throughout the test. Probably for this reason, the chambers occupying the periphery are defined as the "final whorl" and those surrounding the proloculus as the "initial whorl." One obtains this "final whorl" by rotating the test 'backward' 360°, with the final chamber as starting point (Figure 8). These chambers occupy the periph- ery of the test. (Usually, however, the 360° line falls across the whorl's innermost chamber, giving a whorl 180- PERIPHERAL WHORL COUNTING 270- Figure 8.—Peripheral whorl counting. The cross hairs of the goniometer microscope eyepiece are represented by the 0°-180° and 90°-270° lines. They are centered in the proloculus, with the 0°-180° line extending through the center of the proloculus and the outermost basal edge of the final chamber. Measurement starts at the intersection of the line with that basal edge and gives degrees of whorl occupied by each chamber, with a total of 820°. Chambers are num- bered progressively inward, with the final chamber as 1. Each whorl occupies 360°. Therefore, in this case the peripheral whorl is occupied by chambers 1, 2, 3, and most of 4; the pre-peripheral whorl by part of 4, 5, 6, and most of 7; and 720° to 820° by part of 7 and 8. By this method, the peripheral whorl is complete and it is the chambers immedi- ately "younger" than the proloculus which do not complete a whorl. 16 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY containing a number of complete chambers plus a fraction of a chamber.) One obtains the "initial whorl" by rotating the test 'forward' from the proloculus ( Figure 9 ). Continuation of this 'forward' measure- ment usually results in the last complete whorl (360°) extending onto the periphery and thus overlapping the "final whorl." ('Backward' measurement usually gives an "initial whorl" short of 360°.) Although opposite to the direction of growth (see Cifelli, 1961), practical reasons exist for inward meas- urement of chambers with respect to volution. Periph- eral chambers are most clearly visible, becoming progressively less so inward and often barely or not visible in the prolocular whorl. The number and size of chambers occupying the periphery are useful, easily INITIAL/ WHORL COUNTING Figure 9.—Initial whorl counting. The cross hairs of the goniometer microscope eyepiece are represented by the 0°— 180° and 90°-270° lines. They are centered in the proloculus, with the 0°-180° line extending through the center of the proloculus and its intersection with the beginning of the first chamber beyond it. Measurement starts at that intersection and gives degrees of whorl occupied by each chamber, with a total of 820°. Chambers are numbered progressively out- ward from the proloculus. Each whorl occupies 360°. There- fore, in this case, the first whorl is occupied by chambers 1, 2, and most of 3 ; the second by part of 3, 4, 5, 6, and part of 7; and the third by part of 7 and 8. By this method, the third whorl finishes before completing 360° and the chambers around the periphery are not all included in the third whorl. obtained diagnostic criteria. By rotating specimens reverse to growth, one can determine the relationship between chamber number and degrees of whorl throughout much of the test ontogeny, with a fixed point of reference, even though early chambers may be obscure. Therefore, we use this method here. To maintain consistency and avoid ambiguity, however, we propose the term "peripheral whorl" for the sum of chambers that occupy the periphery and comprise a 360° volution when the last-formed chamber is used as starting point. Peripheral whorl is intended to re- place "final whorl" of common usage. Our procedure for counting and measuring cham- bers per whorl follows. The specimen is oriented spiral side upward on a mechanical stage of a binocular microscope having a goniometer eyepiece. The goni- ometer crosshairs next are rotated until the north- south line intersects the suture between the ultimate and penultimate chambers, and the degrees of rotation are recorded. This process continues so far as cham- bers are visible enough to measure. The size (maxi- mum diameter) of each volution (360° of rotation) also is measured and recorded. The choice of intersection point of the north-south line with the final chamber's forward edge and with the previous chamber's sutures offers minor problems. With the final chamber, the point can be at the cham- ber base, where it last touches or lies directly above the earlier-formed whorl, or it can be at the chamber's forwardmost projection. With sutures between pre- vious chambers, the point can be either at their bases or outermost visible extremities. Once these choices have been made, they should be adhered to consistently. We have compiled graphs showing the relationship between chamber addition and degrees of volution for specimens of the species mentioned above. The graphs are arranged "backwards," so to speak, as final chambers are the 0° points, and the curves run op- posite to growth direction. This procedure is necessary to keep the curves open-ended for those specimens whose early chambers are not clearly visible. Further, this method avoids the problem of beginning (and comparing) plots of specimens with their most uncer- tain measurements, of chambers immediately surround- ing proloculi ; even when these chambers can be seen, their measurements may be rather inexact because of their very small sizes. NUMBER 4 Systematic Descriptions Family GLOBIGERINIDAE Carpenter, Parker, and Jones, 1862 Genus Globigerina d'Orbigny, 1826 Globigerina atlantisae, new species Plate 1 : figures 1, 2, 3 Globigerina radians Egger.—Parker, 1958, p. 278, pi. 5: fig. 10. Test compressed, trochospiral, with a rounded, lobate periphery and a rather elliptical outline; chambers rapidly enlarging as added, usually between two and three whorls in the adult with between four and five chambers in the peripheral whorl and from four to five in the pre-peripheral whorl ; number of chambers rang- ing from 8 to 14 in the entire test, usually 10 or 11; chamber shape appearing elongated along the axis of coiling on the spiral side, and perpendicularly to the periphery on the umbilical side, vertically compressed, especially in the peripheral whorl ; sutures distinct, de- pressed, narrow, slightly curved to radial on the spiral side, more curved on the umbilical side, with the spiral suture following around the bases of the chambers of the peripheral whorl and meeting the radial sutures sometimes in a substellate pattern, otherwise lobate; spiral side of test either almost flat or with early whorls raised as a plane above the peripheral whorl ( Plate 1 : figure 2) ; aperture interiomarginal, umbilical-extra- umbilical, a slit reaching close to the periphery, almost covered by the extended final chamber and small flap attached to the base of that chamber; wall finely per- forate, finely hispid, thin; coiling direction both left and right, with left slightly predominant; maximum diameters of primary types 0.23-0.26 mm. Parker (1958) synonymized her form with Globi- gerina radians Egger on the basis of figures for this species given by Rhumbler ( 1909). The original figure is enigmatic, through it resembles Globigerinella aequilateralis superficially. Rhumbler (1909, p. 11) stated that the form figured by him as G. radians was obtained from Egger. Inspection of these figures re- veals, however, that the chambers are not so elongate in the direction of coiling on the spiral side as are those of the present form or Parker's specimens, but instead are distinctly more lobulate. In this regard, they more closely resemble those of some Globigerinita iota Parker. The nature of the figures, although revealing a form 17 similar to both G. iota and Globigerina atlantisae, new species, does not permit truly detailed comparison. Globigerina atlantisae, new species, most closely re- sembles G. quinqueloba egelida, new subspecies. It dif- fers from the latter in its slightly greater curvature of the sutures. This difference is particularly apparent on the umbilical side. The chambers of G. atlantisae are more elongate along the axis of coiling on the spiral side of the test and less spherical (or subspherical) than those of G. quinqueloba egelida. Globigerina atlantisae also has fewer chambers (generally 10 to 11 in the adult test compared to 13 or 14 in G. quinqueloba egelida). In the peripheral whorl there are between four and five chambers instead of the four and a half to five found in G. quinqueloba egelida. Related to this is that there is a consistent difference in the peripheral outline between the two, with an indentation below and adjacent to the final chamber of G. atlantisae which is not present with G. quinqueloba egelida (Fig- ure 10). The aperture of G. atlantisae usually is ob- scured by the final chamber's flap-like extension, while G. quinqueloba egelida shows this feature less com- monly. Further, the aperture of G. atlantisae extends closer to the periphery. Both species are compressed, but G. atlantisae is generally more so, and also generally has a flatter spiral surface. With G. atlantisae speci- mens, chambers tend to occupy more nearly the same number of degrees of whorl (compared with other Globigerina atlantisae Globigerina quinqueloba egelida Figure 10.—Comparison of outlines of Globigerina altantisae, new species, and Globigerina quinqueloba egelida, new sub- species. Chambers of G. atlantisae are more elongate in the direction of coiling on the spiral side. Further, the plan of growth produces a consistant indentation in the outline of specimens, adjacent to the last chamber, whereas G. quin- queloba egelida shows no such indentation; that is, a line drawn tangent to chambers 1 and 4 will not touch chamber 5 in G. atlantisae, whereas it will intersect chamber 5 in G. quinqueloba egelida. 18 3 o 3 o ac a. en 3 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY W- 12 10 - »#•*- >>' ..* PERIPHERAL WHORL '•> » m*•• - 8 * ¦¦>**¦*•'» 16f 44r-'" 6 16 . ;—-' ' 4 i6^4£-:''' r' Globigerina atlantisae is S, (16 SPECIMENS) 2 16 ...•¦¦• " . i . I i i i i i i i i i i lili i i (M 1 1 1 1 1 1 1 1 1 1 i 1 J-----¦-----1----1-----¦ i 1 1-----¦ ¦ 1 1 1-----._1 100 200 300 400 500 600 700 800 NUMBER OF DEGREES OF WHORL 900 1000 1100 Figure 11.—Globigerina atlantisae, new species. A growth pattern expressed as the relationship between number of chambers and number of degrees of whorl for 16 specimens. The measure- ment increase on both axes goes from the final chamber inward, not from proloculus outward. chambers of the same and other specimens) than is the case with G. quinqueloba egélida (Figure 11). Measurements.—Sixteen specimens were selected to determine the relationship between chamber num- ber and degrees of test volution in Globigerina atlan- Table 2.—Sizes and chamber numbers of 16 specimens of Globigerina atlantisae, new species Range Mean Maximum diameter (mm) 0. 16-0. 28 0.22 through peripheral whorl Maximum diameter (mm) 0.05-0. 13 0.09 less peripheral whorl Ratio of diameters of peripheral 3. 2-2. 2 2.4 whorl and rest of test Total number of chambers 8-14 11(11.31) Number of chambers in periph- 4. 2-4. 8 4.44 eral whorl Number of chambers in preper- 4. 4-5. 4 4.98 ipheral whorl (for those 12 specimens continuing beyond this whorl) tisae, new species (Figure 11). The maximum number of chambers recorded is 14, but most specimens contain 12 or fewer chambers. In all but four specimens the chambers occupy more than two but less than three full whorls. The four specimens consisting of less than two full whorls probably represent immature forms. Good consistency exists among specimens in growth pattern and they plot close to a straight line with a relatively small spread of points. The peripheral whorl contains between four and five chambers, which represents a slight decrease in rate of chamber addition from the previous 360° of volution (pre-peripheral whorl). Distribution.—Globigerina atlantisae, new species, is not common in Atlantis II material. It occurs at seven Atlantis 77-13 stations and one Atlantis II-9. It reaches a peak of 2 percent at station 21. Globigerina bulloides bulloides d'Orbigny Plate 1 : figures 5, 6 Globigerina bulloides d'Orbigny, 1826, p. 277, nos. 17, 76.— Brady, 1884, p. 593, pi. 79: fig. 7.—Phleger, Parker, and NUMBER 4 12 - 3 ü g CL CO 3 8 o: uj m 5 < x ü ü. o ce UJ co 19 PERIPHERAL WHORL 10 1«UÉ f »»»*•* 9 m_ I'»»« » • .'Y. . Globigerina bulloides bulloides I (10 SPECIMENS) i i i___i___i_ ' i J__L. 100 200 300 400 500 600 700 800 NUMBER OF DEGREES OF WHORL 900 1000 1100 Figure 12.—Globigerina bulloides bulloides. A growth pattern expressed as the relationship between number of chambers and number of degrees of whorl for 10 specimens. The measure- ment increase on both axes goes from the final chamber inward, not from proloculus outward. Pierson, 1953, p. 11, pi. 1: figs. 3, 4, 7, 8.—Parker, 1958, p. 276, pi. 5; figs. 1-4.—1962, p. 221, pi. 1, figs. 1-8.—Bé, 1959, pi. 1, figs. 15-17.—Bradshaw, 1959, p. 33, pi. 6, figs. 1-4;—Banner and Blow, 1960a, p. 3, pi. 1, figs. 1, 4.— Cifelli, 1965, p. 11, pi. 1, figs. 1-3, 5. Populations of Globigerina bulloides bulloides in the Atlantis II material mainly are easily distinguishable from other groups. Occasional difficulty is encountered in separation from some forms of Globigerinella aequi- lateralis in which the growth plan remains essentially trochospiral. Distinction is made primarily on the apertural position, which in G. aequilateralis is extra- umbilical and tends to extend into the equatorial region. The growth plan of quite immature forms of Globigerinoides conglobatus is almost identical with that of Globigerina bulloides bulloides, but the former species is usually easily distinguishable by its coarsely hispid wall and the occurrence of suppementary apertures. A serious problem of differentiation exists between G. bulloides bulloides and G. bulloides fal- conensis; this is discussed under the latter taxon. One of the few specimens of G. bulloides bulloides with a supplementary aperture is figured (Plate 1; figure 5) because it is unique. Measurements.—Ten specimens were selected to show the relationship between number of chambers and test volution in Globigerina bulloides bulloides (Figure 12). G. bulloides bulloides exhibits a rather distinctive growth pattern among planktonic species, averaging slightly fewer (11) total number of cham- bers and having fewer (between three and four) Table 3.—Sizes and chamber numbers of ten specimens of Globigerina bulloides bulloides Range Mean Maximum diameter (mm) 0. 25-0. 40 0.32 through peripheral whorl Maximum diameter (mm) less 0. 09-0. 19 0. 13 peripheral whorl Ratio of diameters of peripheral 2. 8-2. 1 2.5 whorl and rest of test Total number of chambers 9-12 11. 10 Number of chambers in 3. 5-3. 9 3.8 peripheral whorl Number of chambers in 2.9-4.7 3.9 preperipheral whorl 20 chambers in the peripheral whorl than do many species. All chambers are contained in within two to three test volutions. In the early volutions there is high variability in chamber addition with respect to test volution, but the variability decreases markedly in the pre-peripheral whorl. Distribution.—Globigerina bulloides, sensu lato is represented in all Atlantis 77-13 samples, but seldom in large numbers. Its maximum development is in cold water, especially sample 32, where 29 percent of the assemblage was referred to G. bulloides falconensis. No obvious different distribution pattern exists be- tween the two subspecies, G. bulloides bulloides and G. bulloides falconensis. In Atlantis II-9 samples, G. bulloides, sensu lato dominated at five of the eleven stations. Globigerina bulloides falconensis Blow Plate 1 : figure 4 Globigerina falconensis Blow, 1959, p. 177, pi. 9: figs. 40, 4L—Parker, 1962, p. 224, pi. 1: figs. 14, 16-19. This subspecies tends to be smaller than the subspecies bulloides and has a constricted aperture. Of the re- duced final chamber and more lobulate periphery, considered characteristic by Parker (1962, p. 224) in differentiating Globigerina falconensis from G. bulloides, the latter character appears too variable among the present specimens to be taxonomically significant, and neither seems characteristic of North Atlantic populations here referred to this taxon. Populations from the Pacific referred to G. falconensis by F. L. Parker and kindly showed to the junior author by her, differ from those here referred to G. bulloides falconensis, however. Their chambers do not increase so rapidly in size in the peripheral whorl and thus their outline differs. Two taxa may be represented. Among the present specimens, there are usually between three and four chambers in the peripheral whorl. The earliest peripheral whorl chamber is much smaller than the other three. In the Atlantis II material the subspecies falconensis is completely grada- tional to the subspecies bulloides. Specimens with three chambers almost completely occupying the peripheral whorl closely resemble some immature Globigerinoides trilobus in growth plan. Insofar as we are able to determine from examina- tion of type-specimens, the relatively small forms with SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY relatively constricted apertures associated with and previously included in Globigerina bulloides (see Cifelli, 1965, p. 11, pi. 1: fig. 2) are identical with G. bulloides falconensis of this report. As in our ma- terial these forms intergrade, we propose to treat G. falconensis as a subspecies of G. bulloides. Perhaps it should be emphasized at this point that subspecies as used here refer to populations not to forms or specimens. Most of our populations show considerable variation and include mixtures of bulloides and falconensis, although one or the other usually clearly dominates. We assign populations ac- cording to which form is the most abundant. Thus, Globigerina bulloides falconensis includes those popu- lations where over half of the individuals are recog- nizable as falconensis. In the present material, a major differentiation problem exists between Globigerina bulloides falconen- sis and Globigerinita glutinata, and it is possible that some workers would include many of our specimens in G. glutinata. A gradation occurs between shiny, smooth-surfaced forms with extraumbilical silt-like apertures and three chambers in the peripheral whorl (morphologically like mature Globigerinoides trilobus trilobus but smaller and smoother) and forms with rather opaque, finely hispid walls and with umbilical apertures which, while not as rounded as those char- acteristic of Globigerina bulloides are less silt-like than usual for Globigerinita glutinata. With assemblages from some stations only an almost arbitrary separation is possible between G. glutinata and Globigerina bul- loides falconensis. In these assemblages characteristic Globigerinita glutinata is poorly developed, although some specimens appear to belong to that taxon. It is in- teresting, however, that where G. glutinata is well developed, as at stations 8 and 13, few forms referrable to Globigerina bulloides falconensis occur and the two groups are fairly distinct. Occasionally, differentiating G. bulloides falconensis and Globigerinita glutinata from relatively smooth or shiny immature specimens of Globigerinoides ruber and G. trilobus poses problems. For further discussion of the G. glutinata problem, see page 35. Globigerina rubescens, rare in the present assemblages, has more spherical chambers than G. bulloides falconensis and chambers which in the peripheral whorl, tend to be of approximately the same size instead of increasing markedly in size. The aperture of G. rubescens also is more arched and the average size less than that of characteristic G. bulloides falconensis. NUMBER 4 21 Globigerina dutertrei d'Orbigny Plate 2: figures 1, 2 Globigerina dutertrei d'Orbigny, 1839a, p. 84, pi. 4: figs. 19-21.— ? Brady, 1884, p. 601, pi. 81: fig. 1.—Banner and Blow, 1960a, p. 11, pi. 2: fig. 1.—Cifelli, 1965, p. 12, pi. 2: figs. 1, 2. Globigerina dubia Egger.—Brady, 1884, p. 595, pi. 79: fig. 17. Globigerina eggeri Rhumbler, 1901, p. 19, fig. 20.—Phleger, Parker, and Pierson, 1953, p. 12, pi. 1: figs. 11, 12.— Parker, 1958, p. 277, pi. 5: figs. 5, 7.—Bé, 1959, pi. 2: figs. 1-3.—Bradshaw, 1959 [part], p. 35, pi. 6: figs. 5, 10 [not 8, 9]. Globoquadrina dutertrei (d'Orbigny).—Parker, 1962, p. 242, pi. 7: figs. 1-13, pi. 8: figs. 1-4. Fully developed specimens are distinctive and easily recognizable. They have a large, open umbilical aper- ture, with some development of umbilical teeth and greater number of chambers and degrees of volution than is usual among globigerinids. (The open umbili- cus is found much more frequently among water- column than bottom-sediment specimens. ) Globigerina dutertrei, however, is highly variable and less fully developed forms may closely resemble both G. in- compta and forms here referred to as G. äff. G. pachyderma. The relationship of G. dutertrei with the latter two species still remains obscure and consider- able differences in opinion exist among workers as to synonomies and morphologic limits. Questioned are how many taxa are actually repre- sented by the G. dutertrei-G. incompta-G. pachyderma (or G. aff. G. pachyderma) complex and whether G. incompta is a life stage of either of the other two. From our study of the Atlantis 77-13 and -9 material, we believe that G. dutertrei, G. incompta, and G. aff. G. pachyderma represent three separate taxa. We will present most of our arguments here, under discussion of G. dutertrei. (Bottom-sediment speciments present similar and possibly more complex problems, but they are not discussed per se here. ) Largely, we base our conclusions on differences in development, treated under ontogenies and measure- ments. Because the problem is intricate, some redun- dancy is inevitable, but perhaps will emphasize the reasons for our treatment of these forms. We will start with general remarks on problems and resemblances and differences among these forms. Both Parker (1962, p. 224) and Bé and Hamlin (1967, p. 96) consider G. incompta fully synonomous with G. pachyderma. Bé and Hamlin regard G. incompta as immature G. pachyderma, while Parker considers the differences between the latter two forms and G. dutertrei sufficient to place them in separate genera and subfamilies. (Parker, 1968 personal com- munication, presently is inclined to view G. incompta as more closely related to G. dutertrei and possibly a form of that species. ) Yet, considerable similarity exists between G. incompta and G. dutertrei. When G. incompta was first described, it was compared mainly with G. dutertrei (Cifelli, 1961). A number of forms in the literature identified as G. dutertrei compare fav- orably or are identical with G. incompta. According to Parker ( 1962, p. 242), one of the more distinctive features of G. dutertrei (Globoquadrina dutertrei of Parker) is the "pitted" wall surface. We have noted this "pitted" surface. The "pits" are rather angular and appear to represent open spaces above pores and to be surrounded by discontinuous ridges that look like small papillae joined together. Among the small, ridge-like papillae are a few higher, angular, spinose-appearing projections. For the mo9t part, the ridge-like papillae are continuous around the pore area, but on a single specimen they may range from com- plete continuity, forming a joined network, to a dis- continuous ridge, to isolated, discrete papillae. In Globigerina incompta the papillae tend to be more discrete, and in some cases spines protrude from the angular projections. Some tendency, however, exists for papillae to fuse around the pore area, resulting in a surface texture very similar to that of G. dutertrei. Therefore, distinction between the two species cannot always be made based on surface texture. The wall of G. aff. G. pachyderma is coarser, with a sugary texture, especially in the early stages, and it tends to greatly obscure the earlier chambers. Basically, however, it seems of the same texture, except that the papillae are thicker, more angular, and less regularly arranged. Actually, the fine details of the wall of all species are difficut to ascertain under the light micro- scope. Textural appearance varies with lighting and state of preservation of specimens. Detailed examina- tion with an electron scanning microscope will be necessary to determine the nature of these fine structures. Smaller individuals of G. dutertrei with fewer than average chambers and relatively large final chambers closely resemble, and are sometimes difficult to dis- tinguish from, G. incompta (Cifelli, 1961). We base our separation mainly on the aperture, which tends to be more centrally located and larger in G. dutertrei (from the water column). In G. dutertrei the arrange- 22 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY ment of chambers is less regular and the outline more elliptical. Therefore, we do not believe that G. incompta is an immature stage of G. dutertrei. In the Atlantis 11-13 samples G. incompta was found in large numbers and associated with but few typical forms of G. dutertrei; we believe this supports our contention of two distinct forms. Smaller forms of G. dutertrei also closely resemble fully developed G. äff. G. pachyderma. The latter taxon, however, has a much more constricted aperture that lacks umbilical teeth. The coarse, sugary wall tex- ture and nature of the reduced final chamber, when it is present, also serve to distinguish G. äff. G. pachy- derma. G. dutertrei from the water column may also have reduced final chambers but they do not tend to be bulla-like as do those of G. aff. G. pachyderma. The difference in wall texture between (1) G. aff. G. pachyderma and (2) G. dutertrei and G. incompta is particularly evident in the early stages, as is the dif- ference in amount of depression of sutures, with G. aff. G. pachyderma showing relatively little depression. Ontogenies.—Specimens we believe to represent ontogenetic stages of G. dutertrei, G. incompta, and G. aff. G. pachyderma are illustrated in Figures 13, 14, and 15, respectively. Unfortunately, line drawings can- not reveal adequately ontogenetic differences in wall texture, a feature considered a clue to separation of this complex of species; however, the illustrated speci- mens are in the National Museum of Natural History collections and are available for examination. Early stages of G. dutertrei have thin, mostly smooth, shiny walls with small, discrete papillae. During growth, the wall becomes thicker and coarser. The papillae appear to coalesce, forming the partially continuous ridges encircling the depressions containing the pores. Though degree of coalescence varies, it is not complete, even in the adult. Occassional fully developed speci- mens exhibit a final chamber with a thin, smooth, dis- cretely papillate wall like those of earlier stages of ontogeny. In G. incompta, the very early stages show a wall that is thicker and has larger and more discrete papillae than are usually seen in equally immature G. dutertrei. During growth, the wall of G. incompta tends to follow a similar pattern of coalescence of papillae as occurs in G. dutertrei, except that the coalescence begins earlier in the ontogeny and is less well developed in G. in- compta, leaving many more discrete papillae in the adult form. Thus, the textural pattern of pits and ridges characterizes both G. dutertrei and G. incompta, though it is less pronounced in the latter, and some morphologic overlap occurs. In G. aff. G. pachyderma the ontogenetic textural trend appears just the reverse of that of the two other species. In early stages the wall is extremely thick and coarse, with a sugary texture. The surface is irregular and some spinose and angular papillae can be seen, but for the most part structural details of the texture are difficult to determine. Late in the ontogeny, usually in the last three or four chambers, the wall becomes thinner and the surface more regular. Papillae become more distinct and a pattern of partially coalescent ridges surrounding depressed pore areas can be seen. This pattern bears close resemblance to that of G. in- compta and also some forms of G. dutertrei in which the papillae are less typically coalesced. The G. aff. G. pachyderma wall, however, remains coarser than that of G. incompta, with less regularly shaped papillae that also tend to be more spinose. Thus, while the adult wall of G. aff. G. pachyderma closely resembles that of G. incompta and some forms of G. dutertrei, a distinct difference exists in the early stages and some differ- ences can be seen in the mature forms. The aperture of G. dutertrei closely resembles that of G. incompta in the early stages. Both apertures are at the base of the final chamber and are umbilical- extraumbilical or extraumbilical but do not reach the periphery. In G. aff. G. pachyderma, the aperture is relatively small and consistently umbilical-extraumbili- cal, with the maximum dimension always confined to the umbilical region. During ontogeny, the aperture of G. incompta re- mains constant in both position and relative size. In G. dutertrei (from the water column) the aperture becomes more consistently umbilical and relatively larger and more open early in the ontogeny. This change towards an open umbilical aperture in G. dutertrei relates to ontogenetic coiling change and the relatively small size increase of later chambers, which tend to encircle and maximally disclose the umbilicus. In G. aff. G. pachyderma the aperture remains essen- tially constant during ontogeny except that it becomes more incised because of increased lobulatioh of the later chambers. Thin apertural lips occur throughout the ontogenies of both G. dutertrei and G. incompta. During growth of G. dutertrei, however, some lips develop single, tooth-like lobes (Figures 13d, e, f). In G. aff. G. pachy- NUMBER 4 23 13 Globigerina dutertrei 14 Globigerina incompta 15 d s—^ > *: -¦* ...'¦. V '..• O er °-io - 9* - «i;i * * » • CO ..-:¡®H ¦Wiv:;: 3 Z PERIPHERAL WHORL 10. . .?& ; 2 w 8 - «.....« . CO CE Ld '"•*»« » •« . • m 2 _ 10 < 6 - ¦ "» •*• » X o 10,^ . Ll o 10 : a: 4 %•(.;*.;;.;.¦.•• LlI m ,(Uá 2 »o_______ c Globigerina incompta 2 2 1 t ( 3 r 3 (10 SPECIMENS) __i__i__i__1 i i__i__1__i__i__i—L ID rO ¦ ¦'.....1 , ¦ ¦ 1 ¦ ¦ . 1 ¦ ' ........... . ¦ i i i i_j 100 200 300 400 500 600 700 800 NUMBER OF DEGREES OF WHORL 900 1000 1100 Figure 17.—Globigerina incompta. Growth pattern expressed as the relationship between number of chambers and number of degrees of whorl for 10 specimens. The measurement increase on both axes goes from the final chamber inward, not from proloculus outward. 28 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY Table 6.—Sizes of the ten specimens of Globigerina incompta shown in Figure 17 Range Mean Maximum diameter (mm) 0. 18-0. 33 0.26 through peripheral whorl Maximum diameter (mm) less 0. 06-0. 14 0.095 peripheral whorl Ratio of diameters of peripheral 3. 0-2. 4 2. 7 whorl and rest of test Maximum diameters and chamber numbers of an additional number of specimens from two stations were measured and compared with those of G. duter- trei and G. äff. G. pachyderma (Tables 4, 5). All speci- mens observed are right-coiling. Distribution.—Globigerina incompta is present in almost every assemblage studied. It constitutes less than 10 percent of the assemblages in many samples of the Atlantis 77-13 traverse but reaches 16 percent in sam- ple 4, 32 percent in sample 32, and 64 percent, 72 per- cent, 62 percent, 55 percent, and 44 percent in samples 34, 36, 38, 40, and 42, respectively. In the Atlantis 77-9 samples, G. incompta constitutes more than 10 percent of four assemblages, 6 of one, and less than 5 percent of four. Globigerina inflata d'Orbigny Plate 2: figures 4, 5 Globigerina inflata d'Orbigny, 1839b, p. 134, pi. 2: figs. 7-9.—Phleger, Parker, and Pierson, 1953, p. 13, pi. 1 : figs. 15, 16.—Parker, 1958, p. 277, pi. 6: fig. 3.—Bé, 1959, pi. 1: figs. 12-14.—Bradshaw, 1959, p. 36, pi. 6: figs. 16-18.—Cifelli, 1965, p. 14, pi. 4: figs. 1-3. Globorotalia inflata (d'Orbigny).—Parker, 1962, p. 236, pi. 5: figs. 6-9. This species varies considerably in chamber inflation and apertural character. Normally there are four cham- bers in the peripheral whorl, but some individuals, par- ticularly among small immature ones, contain three or five. Large inflated forms often strongly tend to- wards peripheral rounding and streptospiral coiling in the final whorl. Rarely the final chamber is reduced. The aperture usually is umbilical-extraumbilical but may be extraumbilical or, more rarely, umbilical, and varies considerably in height. Where the last chamber is streptospirally arranged, as proves to be the case with most specimens upon close inspection, the aperture is oblique to the suture between the first two chambers of the peripheral whorl. Some specimens of this type closely approach Pulleniatina in their coiling. In fully trochospiral specimens with four chambers in the peri- pheral whorl, the aperture is approximately perpendic- ular to the suture between the two earliest chambers of the peripheral whorl. Normally, high, large aper- tures are associated with forms with inflated chambers ; lower more restricted (although still large) apertures characterize forms with acute peripheries. As previously noted (Cifelli, 1965, p. 14), the closest morphologic affinities of this species appear to be with Globorotalia punctulata (of Cifelli, 1965), the differ- ence being chiefly in the degree of chamber inflation and concomitant peripheral sharpness and in the aper- tural height. (See Cifelli, 1965 for discussion of the generic placement of Globigerina inflata and related problems.) Although the characters are variable, we have not observed a complete transition between the forms. Complete certainty of identity, however, is not always possible with some small, immature individuals. These tend to have relatively acute peripheries; many also have low, restricted umbilical-extraumbilical aper- tures, although some have rather rounded umbilical apertures. A marked bend in the ventral sutures char- acterizes some of these small specimens. The numbers of these small problematic specimens usually are sig- nificantly smaller than those of more characteristic Globigerina inflata, but in sample 13 they occur almost as abundantly. A comparison of the variational pat- terns between Globigerina inflata and Globorotalia punctulata is shown in Figure 18. Distribution.—Globigerina inflata is well repre- sented in the Atlantis 77-13 traverse assemblages, be- ing present at every station, ranging in relative abun- dances from 2 to 34 percent. The small forms which morphologically also approach Globorotalia punctulata but are taxonomically referred to Globigerina inflata are present in most samples in abundances of from less than 1 to 4 percent. In the Atlantis 77-9 samples Glo- bigerina inflata (including the small form) is very abundant. It is present in all samples, constituting more than 10 percent of every assemblage but one, and 20 percent or greater in all but three, reaching a maxi- mum of 75 percent in sample 288. NUMBER 4 29 Figure 18.—Comparison between and growth series of (g-a) Globorotalia punctulata and (h—t) Globigerina inflata. Note similarity between species of some immature individuals. Globigerina megastoma Earland Plate 3: figure 1 Globigerina megastoma Earland, 1934 (1935), p. 177, pi. 8: figs. 9-12. Two specimens apparently referable to this species were found in the sample from station 5. They are quite distinct from any other forms found in the At- lantis II assemblages. Although not as large (0.34 mm maximum diameter as compared with 0.60 mm given by Earland) as Earland's form from the South Atlantic, these two specimens compare well in having a very thin, fairly smooth wall, a highly trochoid spiral test, chambers greatly inflated and rapidly increasing in size with slightly more than four but not five com- plete chambers in the peripheral whorl, and quite depressed sutures. The nature of the aperture of the present two specimens may differ somewhat from that described by Earland but precise comparison is diffi- cult on the basis of the figures. Earland (1934, p. 177) states that the aperature, "situated on the inner edge of the final chamber is large and semicircular, with a reverted lip." The apertures of the present specimens extend from the umbilical area outward around the periphery to the spiral side of the test. They are large but not semicircular, being rendered rather subrec- tangular in outline by their position and the presence of an outward-protruding lip across their upper margin, which actually compares favorably with Earland's figures. To our knowledge this is the first record of this species in the plankton. It may have been overlooked in the past because of the test's frangibleness. The specimens' appearance, especially the aperture, also suggests that they could be immature individuals of some other planktonic taxon, but none occurring in our material. Globigerina aff. G. pachyderma (Ehrenberg) Plate 3: figures 2, 3 Globigerina pachyderma incompta Cifelli.—Cifelli, 1965 [part], p. 11, pi. 1.: figs. 4, 6. Test trochospirally low, rounded to lobate; chambers subangular to subcircular, numbering approximately 30 SMITHSONIAN CONTRIBUTIONS TO PALEOBIOLOGY 12-13 in test, but so obscured in early volution as to be indeterminate on most specimens, with four to five chambers in peripheral whorl, exact number depend- ing on final chamber size; chambers in peripheral whorl of almost equal size except final chamber which may be reduced; aperture small, umbilical-extraum- bilical but mostly umbilical with slit-like extension towards periphery, often partially overlapped by re- duced final chamber or bulla; wall thick, coarse with surgary texture, hispid, ridged, and pitted; sutures radial, flush in early part to depressed in peripheral whorl, mostly obscure on early chambers; coiling to the right. The specimens included here bear the most re- semblance to Globigerina pachyderma of any we have observed in North Atlantic plankton. The resemblance is closest among the small, not fully developed in- dividuals. These are characterized by extremely com- pact tests with thick walls, which obscure early chambers, by silt-like apertures whose maximum dimensions are centrally located, and by four chambers in the peripheral whorl. Except for the fact that they are on the average smaller, they compare favorably with G. pachyderma observed in bottom sediments from the Arctic and Antarctic regions. The larger, more fully developed specimens, however, become less compact, with the aperture umbilical-extraumbilical and more open, though partly obscured by the down- ward extension of the final chamber. They also usually have a less coarsely surfaced, although still sugary- appearing wall in the latter part of the test. There are between four and five chambers in the peripheral whorl. Where a full fifth chamber appears in the peri- pheral whorl, it is much reduced in size, and sometimes bulla-like. Adult specimens closely resemble and can be confused with G. incompta, as was done by Cifelli (1965, p. 11, pi. 1: figs. 4, 6) with the two specimens he figured as variants of G. incompta. They are dis- tinguishable, however, from G. incompta and G. dutertrei, which they also resemble, mainly on their ontogenetic development which is discussed under G. dutertrei (p. 22). In contrast, a suite of specimens from Pacific bottom sediments kindly sent us by F. L. Parker includes forms comparable to mature Globigerina aff. G. pachy- derma and to mature G. dutertrei. Further, specimens from Atlantic bottom sediments show transition be- tween forms referable to G. dutertrei and G. aff. G. pachyderma, although these forms are not identical to our water-column specimens. No clear division be- tween either the Pacific or Atlantic bottom sediment forms could be made. They seemingly represent tran- sitional series and it might therefore appear that G. aff. G. pachyderma of this report is an immature form of G. dutertrei. Further, the junior author has seen a few specimens at Scripps Institution of Oceanography that were collected from the water column in the Pacific. These specimens are definitely referable to "typical" G. pachyderma and they occur in associa- tion with G. dutertrei. We have no explanation for the apparent differences in relationships of these two water-column groups between the Atlantic and Pacific, nor for the differences between the transitional series from the two oceans. Possibly, G. pachyderma and G. dutertrei are polyphyletic groups. On the basis of the Atlantic water-column material that we have ex- amined, however, the taxa are discrete and we prefer to retain them until relationships are clarified. Perhaps the chief problem with Globigerina pachy- derma is that while it dominates bottom planktonic faunas in the Arctic and Antarctic regions, comparable assemblages mainly are of doubtful occurrence or absent in the plankton. It is possible that they live mainly in water deeper than normally sampled. Ac- cording to Bé (1960a) forms with G. pachyderma morphology do not occur in Arctic surface waters. In G. aff. G. pachyderma of this report, the forms most comparable with G. pachyderma, forma typica, con- stitute but small percentages of the populations. To account for the apparent absence of Globigerina pachyderma, forma typica, in surface waters, Bé con- cluded that the G. pachyderma morphology is achieved after the individuals sink below 200 meters, where addi- tional calcite and a reduced final chamber are added. We cannot reconcile our observations with this con- clusion. First, the forms from Atlantis 11-13 and -9 that are comparable with G. pachyderma come from waters 200 meters or less deep. It may eventually prove biogeographically significant that these waters are temperate, not Arctic. Second, our ontogenetic inter- pretation is opposite to Bé's. We infer that the G. pachyderma found is an early, rather than late, devel- opmental stage. Finally, Arctic plankton assemblages we have observed are composed of a variety of forms of difficult taxonomic placement. We find it hard to see how many of them could assume the G. pachy- derma morphology by Bé's mechanism. NUMBER 4 31 Another interpretation of the G. pachyderma prob- lem, one that deserves serious consideration, is that of Uchio (1960). Uchio suggested that the G. pachy- derma on the Antarctic bottom were deposited some- time before the present and that sedimentation there is slow. Dating of the tests by Cu gave 5,490 (±370) years ago. According to D. J. Stanley (personal communica- tion), the bottom in the region around Nova Scotia, particularly on the slope, and perhaps around the Arctic region as a whole, receives little sediment and the Pleistocene crops out near the surface. There are also indications that the Holocene sediment exposed may be mixed with Pleistocene. Another possibility is that the Arctic and Antarctic cold waters act to dis- solve tests of species with relatively thin walls at depths considerably less than normal calcium carbonate com- pensation depth. Kennett (1966) and Berger (1968), among others, have shown that Arctic and Antarctic bottom waters act particularly aggressively in attacking calcium carbonate. Most of the forms we have ob- served in Arctic plankton have relatively thin walls, so perhaps most dissolve before, or shortly after reach- ing the bottom. In summary, it is worth considering that assemblages of G. pachyderma, sensu stricto, observed on the bottom are extinct or near extinct, and that G. äff. G. pachyderma is a descendant subspecies. Measurements.—The relationship between cham- ber number and test volution of six specimens of Globigerina äff. G. pachyderma is shown in Figure 19. The coarse wall surface obscures chambers in early volutions and the exact total number clearly shows on only two of the six specimens. For this reason, queries are placed at the initial volutions in Figure 19. The recorded maximum number of chambers on any specimen is 14, but probably the actual maximum is not much greater, since the wall totally obscures only a minute portion of the test. Chamber addition in this species appears relatively consistent throughout the observable part of the 14- 312 3 o CE °-io CO ID (fig.), 13, 14, 15, 39, 40, 50 (pi.) Rotalina, 39 humilis, Globigerinita, 7 (fig.), 36,49 (pi.) Truncatulina, 36 incompta, Globigerina, 5, 6 (fig.), 7 (fig.), 8 (fig.), 9, 10, 11, 12, 13, 14, 15, 21, 22, 23 (fig.), 24, 25, 26, 27 (fig.), 28, 30, 32, 34,46 (pi.) 51 52 INDEX inflata, Globigerina, 5, 6 (fig.), 7 (fig.), 8, 9, 10 (fig.), 13, 14, 15, 27, 28, 29 (fig.), 40, 46 (pi.) Globorotalia, 28 iota, Globigerinita, 17 mediterranensis, Planorbulina, 7 (fig.), 13,42 megastoma, Globigerina, 7 (fig.), 29,47 (pi.) menardii, Globorotalia, 7 (fig.), 40 Pulvinulina, 40 Rotalina, 40 michelineana, Pulvinulina, 41 murrayi, Hastigerina, 41 Nonionina, 41 pelágica, 41 Orbulina, 42 suturalis, 42 universa, 7 (fig.), 14, 42,50 (pi.) pachyderma, Globigerina, 21, 26, 27, 30, 31, 33 äff. G. pachyderma, Globigerina, 7 (fig.), 15, 21, 22, 23 (fig.), 24, 25, 26, 28, 29, 30, 31 (fig.), 32, 47 (pi.) pachyderma incompta, Globigerina, 26, 29 patagónica, Pulvinulina, 41 pelágica, Hastigerina, 7 (fig.), 36, 41, 42 Nonionina, 41 Planorbulina, 42 mediterranensis, ? (fig.), 13,42 Pulleniatina, 28 Pulvinulina, 39, 40, 41 canariensis, 39 crassa, 40 menardii, 40 michelineana, 41 patagónica, 41 sei tula, 41 punctulata, Globigerina, 40 Globorotalia, 7 (fig.), 28, 29 (fig.), 39, 40, 50 (pi.) quinqueloba, Globigerina, 32, 33 äff. G. quinqueloba, Globigerina, 32, 33 cf. quinqueloba, Globigerina, 32, 33 quinqueloba egelida, Globigerina, 5, 6 (fig.), 7 (fig-), 8 (fig.), 9, 10, 11, 13, 15, 17 (fig.), 18, 27, 32, 33, 34 (fig.), 47 (pi.) radians, Globigerina, 17 Rotalina, 39, 40, 41 hirsuta, 39 menardii, 40 truncatulinoides, 41 ruber, Globigerinoides, 5, 6 (fig.), 7 (fig.), 8 (fig.), 9, 10, 11, 13, 15, 20, 35, 36, 37, 38 (fig.), 39, 49 (pi.) rubescens, Globigerina, 7 (fig.), 20, 35, 48 (pi.) rubra, Globigerina, 38 Globigerinoides, 38 sacculifer, Globigerinoides, 39 sacculifera, Globigerinoides, 38 scitula, Globorotalia, 7 (fig.), 41, 50 (pi.) Pulvinulina, 41 siphonifera, Globigerinella, 35 suturalis, Orbulina, 42 Tretomphalus, 42 atlanticus, 7 (fig.), 13,42 triloba, Globigerina, 38 trilobus, Globigerinoides, 20, 36, 38, 39 trilobus trilobus, Globigerinoides, 7 (fig.), 20, 36, 38, 39 (fig.), 50 (pi.) Truncatulina, 36 humilis, 36 truncatulinoides, Globorotalia, 5, 7 (fig.), 13, 14, 41 Rotalina, 41 universa, Orbulina, 7 (fig.), 14, 42, 50 (pi.) U.S. GOVERNMENT PRINTING OFFICE: 1970 O—359-866