—_ =>" = = Oe bat Wy PPh) Say oe di, ; 1 he Jara ir silesk aba af “y a) ene ek. i Bh s “9 ) a BULLETIN OF THE eee Harvard Lhiverstty . — MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, IN CAMBRIDGE. VOL. XXXVI. - CAMBRIDGE, MASS., U.S. A. 1900-1901. QL ee KS \) 36-3) University Press: Jouyx Wrirson anp Son, Campriner, U.S. A. CONTENTS. No. 1.—An Arvantic “ PaLoLo,” SraAUROCEPHALUS GREGARICUS. By ALFRED GoLpsBorouGH Mayer. (3 Plates.) June, 1900. No. 2.— Some Norrn American Fresu-WareR RuoYNCHOBDELLID&, and their parasites. By W.E.Casrie. (8 Plates.) August, 1900 No. 3.— Fossiz Leprposrerps from the Green River Shales of Wyoming. By C. R. Eastman.” (2 Plates.) August, 1900 . No. 4. — Characters and Relations of GALLINULOIDES, a Fossil Gallinaceous Bird from the Green River Shales of Wyoming. By Freperic A. Lucas. (1 Plate.) August, 1900 No. 5.— The DeveLorment of the Mourn-Parts of ANURIDA MARITIMA Guér. By Justus Watson Fotsom. (8 Plates.) October, 1900 ‘No. 6.— Reports on the Dredging Operations off the West Coast of Central America to the Galapagos, to the West Coast of Mexico, and in the Gulf of California, in charge of ALEXANDER AGaAssiz, carried on by the U. S. Fish Commission Steamer “ Albatross ” during 1891, Lieut. Commander Z. L. Tanner, U.S.N., Commanding. XXVIII. Description of two new Lizarps of the genus ANoxis from Cocos and Malpelo Islands. By Lron- HARD STEJNEGER. (1 Plate.) November, 1900 . No. 7.— Contributions from the Zodlogical Laboratory of the Museum of Comparative Zodlogy at Harvard College, under the direction of E. L. Mark. No. 123. The Orocyst of Decarop Crustacea: Its Structure, Development, and Functions. By C. W. Prentiss. (10 Plates.) July, 1901. . : P No. 8.— Ona Coxtection of Birps from the Liu Kiu Islands. By Ourram Baynes. July, 1901 PAGE 15 65 ig 159 165 253 re Ae Digitized by the Internet A in 2010 with funding from University of Toronto http://www.archive.org/details/bulletinofmuseu3¢ Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE. Vou. 2 AA4Vi. “No, 1, AN ATLANTIC “PALOLO,” STAUROCEPHALUS GREGARICUS. By ALFRED GOLDsBoROUGH MAYER. Witn THREE PLATES. : CAMBRIDGE, MASS., U.S. A.: | PRINTED FOR THE MUSEUM. i June, 1900. No. 1.— An Atlantie “ Palolo,” Stawrocephalus gregaricus. By ALFRED GOLDSBOROUGH MAYER. Durine the summers of 1898 and 1899 I was acting as assistant to Dr. Alexander Agassiz in making a study of Medusz at Loggerhead Key, one of the Tortugas Islands, Florida; and it was while thus en- gaged that the remarkable breeding. habits of the worm about to be described were observed. It gives me pleasure to eXpress my appreciation of the generous kind- ness of Dr. Agassiz, to whose permission I owe the privilege of publish- ing this paper. It is also a pleasure to remember the constant interest and kindness of George R. Billbury, Esq., head keeper of the lighthouse at Loggerhead Key, who did everything in his power to further the scientific work, and to render my stay at the Tortugas enjoyable. I also wish to thank Major J. E. Sawyer, U. S. A., who kindly allowed the use of the government steamer, ‘“ George W. Childs,” in transporting me and my apparatus to and fro from Key West to the Dry Tortugas. The worm about to be described in this paper appears to possess breeding habits so closely similar to those of the well-known Palolo worm ! of the South Pacific that I am inspired’ to give to it the title of the Atlantic “Palolo.” Our Atlantic “ Palolo,’”? however, is a new species of the genus Staurocephalus, and is therefore quite distinct from the Palolo or Bololo worm (Palolo viridis, Gray ; Lysidice viridis, Collin) of Samoa and Fiji, that swarms in vast numbers, for breeding purposes, upon the surface of the ocean, early in the morning of the days of the last quarter of the October and November moons. 1 It is not the purpose of this paper to discuss the habits of the Pacific Palolo. Good scientific accounts of its wonderful swarming habit may be obtained from the writings of S. J. Whitmee, 1875; W. C. McIntosh, 1885; A. Collin, 1897; B. Friedlander, 1898; and A. Agassiz, 1898. See “ Bibliography” at the end of this paper. VOL. XXXvI.— No. I. 1 7 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. We will first present an account of the swarming of the Atlantic Palolo, and will then give a description of the adult worm, a history of the development of its larva, and finally some general conclusions con- cerning the breeding habits of Polychete. It seems probable that the time of the swarming of the Atlantic Palolo is directly related to the date of the last quarter of the moon, for in 1898 the swarm occurred on July 9, and the last quarter of the moon on July 10; while in 1899 the worm swarmed on July 1, and the last quarter of the moon fell on June 29. In 1898 about two hundred specimens of the worm were seen to swarm on the morning of July 8, but on the following day the animals appeared in vast numbers, while on July 10 only about a dozen specimens could be found after a careful search. In 1899 a wonderfully dense swarm appeared suddenly on the morning of July 1, and only a few worms were to be seen on July 2, after which they disappeared. As it was my habit to sail out upon the ocean early every morning, I am certain that no other swarms than the above-mentioned ones occurred between June 25—August 19, 1898; and May 17-July 4, 1899. Description of the Swarming. — The swarming commenced very early in the morning before sunrise, and soon vast numbers of the worms were to be seen swimming upon the surface of the ocean. Few or none of them were to be found in the shallow water near the shore of Log- gerhead Key, but at some distance to the westward of the island, where the water was between two and five fathoms in depth, they appeared in astonishing numbers. The bottom at this place is of coral-sand, and is covered with a sparse growth of Gorgonians and Nullipore Algz, while nearer the shore the bottom consists of living coral and coral-rock with but little sand. When first observed, at four o’clock in the morning of the days of the great swarms, the worms presented very much the ap- pearance shown in Figure 1, Plate 1. They swam with great activity and as near as possible to the surface of the sea. I estimate that there may have been about two worms to each square foot of the ocean’s surface. The worms were not uniformly distributed, however, but were scattered irregularly, sometimes congregating momentarily in wriggling masses, such as were likened by Agassiz, in the case of the Fijian ” These congregations are not due to Palolo, to “ thick vermicelli soup. any affinity for one another on the part of the worms, but are merely the result of accident, for each individual worm swims about quite inde- pendently of the others, and shows no tendency to remain in the presence of any other which it may chance to meet in its wanderings. The Fe mg MAYER: STAUROCEPHALUS GREGARICUS. 3 worms continued to increase in numbers until the time of the rising of the sun, and then, as the light of the early morning fell upon them, a series of contractions came over the sexually ripe segments of each worm and the eggs or sperm were thus discharged into the water (see Figure 2, Plate 1). This contraction is often so sudden and so violent that the ripe segments are torn asunder, at short intervals, by the breaking of the cuticula, forming large rents through which the genital products escape. The 25-30 anterior segments of the worm contain no sexual elements, and take no part in the contraction, so that they re- main uninjured, and always retain their natural shape and appearance. After the discharge of the sexual products the worms continue to swim near the surface for a considerable time, dragging their torn and contracted sexual segments after them. Sometimes, indeed, the con- traction causes the sexual segments to break away from the anterior portion of the worm, and they swim about, apparently suffering no in- convenience, although without a head. After the discharge of the eggs or sperm the sexual segments become very brittle, and a touch of the hand is often sufficient to cause them to break suddenly into small frag- ments. It seems not improbable that the torn and contracted sexual segments may eventually slough off from the 25-30 anterior ones, and that thus the life of the individual may be saved to perpetuate the species. This, however, is mere conjecture upon my part, for in 1898 all of the worms which were confined in aquaria died during the course of the day without having thrown off their dishevelled posterior seg- ments ; and in 1899, when four of the worms were placed in a large aquarium the bottom of which was covered with sand and stones, three of the worms crawled under the stones, but all died within two days without having thrown off their contracted sexual segments. At 6.30 a.m. the worms began to sink down upon the sandy bottom of the ocean, and by nine o’clock in the morning none of them were to be seen. Large numbers of fish devour the worms during the time of swarming. There is little or no sexual color difference in the worms, both males and females being dull brick-red. The females, however, are sometimes of a duller and more yellowish tint than the males. The sperm is yel- low-buff or slightly pink in color, while the eggs are yellow or greenish yellow. The genital products escape in such quantity that the sea is rendered milky over wide areas, and long after the worms have disap- peared the eggs remain floating near the surface in visible windrows of yellowish color. + BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. In 1900 the last quarter of the moon occurs on June 19 and July 18; and as we do not yet know the limits of the lunar month in which the worm swarms, we may look for it within three days of either of the above dates along any of the Bahama or Florida reefs. It seems not improbable that it swarms annually on one day of the year, and that this day falls within three days of the moon’s last quarter in the month extending from June 15 to July 15. Description of the Adult Worm.— The genus Staurocephalus was founded by Grube, 1855, who has given a synopsis of the genus and a description of all of the then known species in the Jahres-Bericht der Schles. Gesell. fiir vaterl. Cultur., Bd. 56, pp. 109-115, 1878. Since then two new species have been described by McIntosh (’85, pp. 231- 235) ; and references to previously described species have been given by Ehlers, Verrill, and Andrews. Generic Characters. — Annelida, Polycheeta, Family Nereide ; body vermiform, segments distinct. The head-lobes give rise to one or two pairs of jointed tentacles. When two pairs of tentacles are present, one pair arises from the side, and the other from the ventral surface. Eyes are sometimes present. The two first segments are without parapodia. The parapodia possess dorsal and ventral cirri. The dorsal cirrus is often unjointed, but sometimes possesses a short terminal segment. The ventral cirrus is shorter than the dorsal and is unsegmented. The posterior segment has two long dorsal and two short ventral cirri. The upper jaw consists of two simple, connected pieces. The lower jaw consists of two rod-like pieees which approach each other near the middle but diverge both in front and behind. (See Figures 20, 22, 26, 27, Plate 3.) Specific Characters ; Adult Worm. —The worm is about 120-150 mm. in length; and may be even longer, for the posterior segment has not been observed. The segments are distinct, and there are about 17 raetameres per centimetre of the worm’s length. The worm is about 4mm. broad. The ventral surface is quite flat and a deep groove runs down its centre. The dorsal surface is arched, and the dorso-ventral diameter is about 3mm. There are no eyes, but the hypodermis cells of the front end of the prestomium bear a dark rosin-colored pigment, the presence of which may indicate a general sensibility to light. There are no lateral tentacles upon the head, but the ventral prestomium gives rise to two quite stiff tentacular cirri (see Figures 1-3, 9-12). These cirri consist each of but a single joint. An axial nerve runs down the centre of each tentacle, and this nerve is surrounded by Sediceens ghia ae MAYER: STAUROCEPHALUS GREGARICUS. 5 elongate hypodermis cells. The first metamere back of the head ugually bears a pair of very rudimentary parapodia, each consisting of but a short dorsal and ventral cirrus. (Figures 11,12.) In the worm shown in Figure 3, Plate 1, the first three segments back of the head bear very minute and undeveloped parapodia. The parapodia of the body segments are all similar each to each and consist in a well-developed dorsal cirrus, a central setigerous lobe, and a ventral cirrus that is shorter than the dorsal. (See Figure 13, Plate 2.) The setigerous lobe bears four kinds of seta. Most dorsal of all are three or four long curved, slender bristles having a delicately serrated edge (a, Figure 4, Plate 1). Immediately below these there are three or four smaller and more slender bristles, having flat spatula-shaped distal ends that exhibit sharp serrations (+, Figure 4). The ventral half of the setigerous lobe bears five or six sete of the sort shown in d, Figure 4 ; and most ventral of all there is a single thick, stiff bristle c, Figure 4. The blood of the worm is red, and there is a large red-colored blood sinus at the base of the dorsal cirrus of each parapodium. (See Figure 13.) The 25-30 anterior segments contain no sexual elements, these being found, however, in all of the more posterior segments. The blood vessels and nephridia of the sexually mature segments are much larger than are the corresponding organs in the anterior segments. The nephridia of the sexual segments evidently serve to carry off the eggs or sperm. The nephropores (xp, Figure 13) are found at the base of each parapodium near the ventral surface. Sections of the worm were made, but the histology is so closely similar to that of other well-known Nereidée that we consider it unnecessary to enter into details concerning it. The constriction of the sexual segments is due to the powerful contraction of the circular muscles that lie immediately-beneath the hypodermis. The sexes are separate, and there is no distinctly marked sexual coloration. The general color of the worm is dull brick-red or ochre-red, and there is a row of diamond-shaped dull white spots, one in each metamere, running down the mid-dorsal line (see Figure 10, Plate 2). Dark brown pigment is found around the orifice of each nephridium (np, Figure 13), and there are some indistinct brownish spots on the ventral side of the head (see Figure 12, Plate 2). These are not found, however, in all individuals, and probably do not function as eye-spots. Development. — The eggs and larve were killed in Perenyi’s fluid, stained in Kleinenberg’s hematoxylin, imbedded in paraffin and sectioned, the sections being usually of about 6.6 uw in thickness. 6 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. After expulsion from the body of the worm the eggs float near the surface, where they are immediately fertilized. The eggs are quite large ; measurements of the embryos in the 16-cell stage gave the diameter 0.36 mm. The segmentation is total and unequal. Four large yolk-laden macromeres are cut off from the four smaller yolkless micromeres. These latter then divide repeatedly and overlap the four macromeres, and thus the gastula is formed by epibole. Although my observations are far too incomplete for anything but general conclusions, it appears that the early stages of the segmentation are strikingly similar to those of Nereis as described and figured by Wilson (92). Figure 5, Plate 1, represents an embryo in the 16-cell stage, which occurs about three hours after extrusion into the water. It will be seen that the large macromeres are heavily laden with deutoplasm-spheres, while the protoplasm of the micromeres is finely and uniformly vacuo- lated, giving the appearance, when seen in sections, of a delicate network. The centrosomes are of large size and stain quite deeply in haematoxylin. Figure 6 represents the condition of an embryo 93 hours old in which the blastopore (bp) is just about to be closed. It will be seen that a distinct segmentation cavity (sgc) makes its appearance at this stage. This cavity may, however, be due to the action of reagents, and may not represent the natural condition. Unfortunately, all of my material having been killed in Perenyi’s fluid, I am unable to make any state- ments concerning this point. It will be noticed that some of the micromeres at this stage are beginning to exhibit large intracellular vacnoles. This is especially true of those cells about 180° away from the blastopore, and also of some in the immediate vicinity of the blastopore. In later stages this vacuolization affects all of the cells of the embryo, both those of the ectoderm and entoderm, and it is certainly true that for the first week of its life the larva owes its increase in size almost entirely to the remarkable development of intracellular and intercellular vacuoles. In this connection it is interesting to note that Davenport (97) has shown that in the case of tadpoles the early growth is almost entirely due to the imbibition of water. Soon after this, when the embryo is about 94 hours old the blastopore closes, and the large deutoplasm-laden cells are completely enclosed by the micromeres. ‘The embryos then become uniformly ciliated and swim about with consider- able rapidity. Figure 14, Plate 2, represents an embryo 24 hours old. Two eye- spots are now beginning to appear, and between these there is a col- lection of greenish-colored cells. These cells stain very deeply in MAYER: STAUROCEPHALUS GREGARICUS. 7 Kleinenberg’s hematoxylin, and appear to be filled with a mass of deeply stained granules that may represent the coagulum of some fluid. Figures of these cells, in older larve, are shown in(g/) Figures 7, 8, Plate 1. I believe them to be glands, and they are probably homologous with the “frontal bodies” found by Wilson (’92, p. 421) in the larva of Nereis, and perhaps also with the “ problematic bodies” observed by Mead (97, p. 256) in the larva of Amphitrite. Malaquin (93, p. 395, Plate XIV., Figures 12-16) has also found glands in a similar position in the head of the larva of Autolytus Edwarsi. Figure 15, Plate 2, represents a larva 3} days old, and Figure 7, Plate 1, shows a dorso-ventral section of the same. The eyes are now quite large, and the green patch representing the gland cells is very prominent. There are now three bands of cilia: a broad oral band, a narrow post-oral, and an anal band. Two, sets of setz, consisting each of three bristles, have made their appearance immediately posterior to the post-oral band of cilia. These setz originate in folds of the hypodermis. A longitudinal dorso-ventral section (Figure 7) of the worm in this stage shows the very large gland cells (g/) of the head. The mouth (m) shows signs of being about to break through, although as yet it is not functional. The same may perhaps be said of the anus (an). The mid gut (st) of the worm now consists of a delicate ento- dermal epithelium enclosing a mass of highly vacuolated cells laden with yolk spheres. Figure 16, Plate 2, shows a larva 5} days old, and Figure 17 illus- trates the character of the sete from the same worm. Most dorsal of all there is a single long seta (Figure 17, 6) and immediately be- low this there are two setz of the sort shown in Figure 17, a. Figure 18, Plate 3, shows a larva 10 days old. The worm is now 0.5 mm. in length, and possesses three sets of setze. Until the end of the 15th day the larve are remarkable for exhibiting a strongly posi- tive phototaxis. They swim through the water at all depths, but large numbers of them are sure to be found clustered together in those parts of the aquaria where the light is strongest. At the end of the 15th day the cilia disappear, and the worms cease to swim through the water, and sink to the bottom. Figures 19, 20, represent a young worm that is 16 days old, and Figure 8, Plate 1, shows a dorso-ventral longitudinal section of the same. There are now four pairs of parapodia provided with dorsal and ventral cirri. A number of sensory hairs are found scattered over the preestomium, and the posterior segment of the body exhibits a pair of dorsal cirri. 8 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. The mouth opens on the ventral surface, and a dorsal and ventral pair of “teeth” have made their appearance in the cesophagus (see Figure 20). The worms are now about 0.8 mm. in length. Internal as well as external evidences of segmentation now appear (see Figure 8, Plate 1) and the dissepiments (ds) are complete. The walls of the mid gut are very thick and consist of large, irregularly shaped, highly vacuolated cells containing a number of yolk spheres. The cells of the cesophagus (oes) are of an epithelial character. The peripheral circular muscles and the deeper lying longitudinal muscle strands are beginning to appear, and the ventral nerve chain (n) is very apparent- In fact, the animal is no jonger a larva, but is a young worm. Figures 21-23, 25-27, illustrate the condition of the worm at the end of the 26th day. There are now five pairs of parapodia, and the dorsal and ventral cirri of the posterior segment have become long and prominent (see Figure 23). The dorsal and ventral jaws of the cesoph- agus are shown in the side view of the head given in Figure 22. Figures 26 and 27 are views from above and from the side, respectively, of the dorsal pair of jaws. The condition of the ventral pair of jaws is still quite similar to that in the 16-day-old worm shown in Figure 20. The worms are now 1.2 mm. long. They burrow readily beneath the surface of sand, but never swim through the water. Figure 24 shows the condition of a worm 34 days old. The animal is now 1.5 mm. in length, and there are still only five pairs of para- podia. The mature coloration is beginning to appear in two reddish- colored spots immediately back of the eyes. I did not succeed in rearing any worms beyond this condition, and know nothing of the mode of formation of the prestomium and cephalic cirri of the adult worm. It will be observed that in the young worm the mouth opens on the ventral surface and the prestomium is supra-oral, while in the adult worm the prestomium and cephalic cirri are sub-oral (compare Figures 3 and 22). General Conclusions. Remarkably little has been written concerning the egg-laying habits of Polychetae. Wilson (’92, p. 371) states that the eggs of Nereis limbata and N. megalops are discharged at night while the animals are swimming upon the surface of the water. The egg-laying season extends at least from June until September. “The animals appear in abundance only on warm still nights, and even then are rarely found MAYER: STAUROCEPHALUS GREGARICUS. 8) unless the water has been quiet for some days.” ‘“ When the con- ditions are favorable, they come forth soon after dark and swim rapidly about at the surface, sometimes in almost incredible numbers.” It would probably be advantageous to any species of worm already possessed of some such egg-laying habits as those of Nereis to have the duration of the egg-laying period restricted to as short a time as pos- sible, and also to have it occur in that part of the year most favorable for the safety and development of the larva. With equal numbers of mature individuals of two species (a) and (6), if (a) possess a long egg-laying period and (4) a short one, there will be more individuals of (b) discharging sperm or ova at any given moment than there will be of (a), whose breeding season is longer. Consequently the eggs of (6) will be more certain of fertilization, other things being equal, than those of (a). For example, if V represent the total number of individuals of species (a), and also of species (2) ; and if 7’ represent the duration of the egg-laying period of species (a) and ¢ that of species (4): then in any definite unit of time there will be individuals of species (a) . F ; Po eee discharging sperm and ova, while at the same time = individuals of species (5) will be engaged inthe same act. Consequently, if the areas of the breeding-grounds of the two species are equal, there will be ae nae 7 times as many individuals of species () discharging sperm or ova at any moment, in a unit of area, than there are of species (a) engaged in the same act. Then in an area containing m individuals of species (a) there are = individuals of species (>). Therefore the /™ 7 a ae Vm —1 times as far as in the case of species (0). Hence the spermatozow of /m — 1 species (5). We sce, then, that a shortening of the egg-laying season causes a greater concentration of breeding individuals, and therefore shortens the average distance that the spermatozoa must travel in order to fertilize the ova; and as spermatozoa cannot survive for any great length of time, this is an advantage to the species. In this connection average distance apart of the individuals of species (a) is species (a) will be obliged to travel times as far as those of 10 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. it is interesting to notice that according to Wilson (92, p. 372) the males of Nereis outnumber the females to a very remarkable degree, while in Staurocephalus gregaricus, and in the Pacific Palolo, the males and females are about equal in numbers each to each. It is most essential for the perpetuation of the species that the fertilization of the ova should be insured. A very few males placed near to the females will insure this; but where the egg-laying period is a long one, and there are not often great concentrations of individuals, the males must out- number the females in order to make certain that the ova of any given female may be fertilized. The egg-laying period of Staurocephalus gregaricus occurred in 1898 and 1899 upon days very close to the day of the last quarter of the June-July moon. At this time, in the Tortugas, Florida, the summer is well established, the trade winds are no longer steady or boisterous, and the calm weather that precedes the hurricane season has set in. It is interesting to notice that very similar meteorological conditions prevail in Samoa and Fiji, in October and November, — the months of the swarming of the Palolo. My friend, Dr. Charles B. Davenport, has called my attention to the fact that the advantages derived from a short egg-laying season are in some measure offset by the circumstance that under such conditions a large number of young larve are suddenly produced, and that therefore the struggle for food must be greatly increased. To counterbalance this difficulty, however, we have the interesting fact that while the eggs of Nereis contain but little yolk, the eggs of Staurocephalus gregaricus are heavily laden with yolk material. When we learn more concerning the egg-laying habits of Annelids, there will no doubt be a number of species found that possess such swarming habits as those of Nereis, and perhaps a few may be dis- covered in which the breeding season is as short as in Staurocephalus gregaricus and Palolo viridis. In 1893, while acting as assistant to Dr. Alexander Agassiz upon the “ Wild Duck” Expedition to the Bahama Islands, I had the opportunity of observing the swarming of an Annelid, We were anchored off Watlings Island (San Salvador) on the night of January 15, and in Clarence Harbor, Long Island, on the night of January 16. On both of these nights the surface of the sea was covered by thousands of little Annelids. They were translucent, and had large red eyes. They appeared to be congregating for breeding purposes, and were breaking into pieces, so that we often found fragments 50 mm. in length swimming about without a head, The last quarter of the moon MAYER: STAUROCEPHALUS GREGARICUS. 11 occurred on January 9, 1893, and their swarming probably had no relation to this event. Among worms, where certain segments of the body became sexually mature while others remain immature, or non-sexual, we find an inter- esting series of gradationsin complexity. Beginning with Staurocepha- lus gregaricus, where the sexual and non-sexual segments are exactly alike in external appearance, and where the entire worm swims at the surface at the breeding period, the next advance in complexity is met with in Palolo viridis, where, according to Friedlander (1898) the non- sexual segments are very different in appearance from the sexual, and where the sexual segments break off from the anterior portion of the worm and swim about during the egg-laying period without a head. Most complex of all are the cases of Autolytus, Filigrana, Myriana, Procereea, Syllis, etc. (see A. Agassiz, 62; Malaquin, 93, etc.), where the sexual segments acquire a head, and eventually become free swim- ming worms, thus producing an alternation of generations. It seems probable that Staurocephalus gregaricus and Palolo viridis have independently acquired quite similar breeding habits through the agency of similar influences of natural selection ; although it must still be admitted that there remains a possibility that both worms may have descended from a remote and common ancestor that possessed some such breeding habits. The following table will serve to illustrate the principal points of relationship in the breeding habits of the two worms :— THe Atuantic “ PaLoto.” THE PaciFic PALoLo. Staurocephalus gregaricus, Mayer. On July 9, 1898, and July 1, 1899, the worm swarmed in vast numbers, for breeding purposes, at the Dry Tortugas Islands, Florida. The last quarter of the moon occurred on July 10, 1898, and June 29, 1899. The 25-30 anterior segments of the worm contain no sexual ele- ments, the eggs or sperm being found in the posterior body seg- ments. The anterior segments, how- Palolo viridis, Gray, 1847. Lysidice viridis, CoLti1n, 1897. The worm swarms in great num- bers, for breeding purposes, at Samoa and Fiji, upon the mornings of the day of, and the day preceding, the last quarter of the October and No- vember moon. (See Whitmee, 1875; Friedlander, 1898.) According to Friedlander, 1898, a number of the anterior segments of the worm contain no sexual elements, these being found in the posterior body segments. The anterior seg- 12 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. ever, are similar to the sexually developed posterior ones in external appearance. The entire worm swims at the sur- face during the breeding period. The eggs or sperm are extruded from the sexual segments by a series of contractions. They pass out into the water not only through the ne- phridial openings, but also through rents and tears in the body wall of the worm, which are often produced by the violence of the contractions. This action usually occurs soon after sunrise. There is no well-marked sexual color difference, both males and fe- males being brick-red, or ochre-red. The eggs are greenish-yellow and the sperm buff-pink. The males and females are about equal in number each to each. The segmentation is total and un- equal, and the gastula is formed by epibole. The larva is telotrochal. The sete appear very early in devel- opment. The larva possesses a pair of eyes, and remarkably large ecto- dermal, cephalic glands. Harvarp University, April, 1899. ments are of greater breadth and less length than are the sexually devel- oped posterior segments. (See Fig- ure by Friedlander.) The posterior or sexual segments, only, swim at the surface during the breeding period. The anterior por- tion of the worm remains below. The eggs or sperm are extruded from the sexual segments by a series of violent contractions. They pass out into the water not only through the nephridial openings, but also through rents and tears in the body wall of the worm, produced by the violence of the contractions. This action usually occurs soon after sun- rise. (See McIntosh, 1885 ; A. Agas- siz, 1898.) The males are brown, and the fe- males dark green. The eggs are green. (See Whitmee, 1875; McIn- tosh, 1885.) The males and females are about equal in number each to each. The development is unknown. MAYER: STAUROCEPHALUS GREGARICUS. 13 BIBLIOGRAPHY. Agassiz, A. °>62. On Alternate Generations in Annelids, and the Embryology of Autolytus cornutus. Boston Journ. Nat. Hist., Vol. VII. pp. 384-409, Plates IX.-XI. Agassiz, A. °98. Islands and Coral Reef of the Fiji Group. American Journ. Sci., Ser. 4, Vol. V. p. 128. Andrews, A. E. 91. Report on Annelida Polycheta of Beaufort, North Carolina. Proc. U. 8. National Museum, p. 288. Collin, A. 197. Ueber den Bau der Korallenriffe, ete., von Dr. Augustin Kramer, Marinearzt, nebst einem Avhange: Ueber den Palolowurm von Dr. A. Collin. Kiel und Leipzig, Verlag von Lipsius und Tischer. Davenport, C. B. 797. The Réle of Water in Growth. Proc. Boston Soe. Nat. Hist., Vol. XXVIII. No. 3, pp. 73-84. Ehlers, E. 6468. Die Borstenwiirmer, pp. 422-442. (Staurocephalus.) Ehlers, E. °87. Florida-Anneliden. Mem. Mus. Comp. Zodl. Harvard Coll., Vol. XV. pp- 64, 68. (Staurocephalide. ) Friedlander, B. 98. Ueber den sogenannten Palolowurm. Biolog. Centralblatt, Bd. XVIII. pp. 337-357, 2 Figs. Gray, J. E. 47. An Account of the Palolo, a Sea Worm eaten in the Navigator Islands. Proe. Zoél. Soc. London, pp. 17-18. Grube, A. E. 55. Archiv fiir Naturgesch., Jahrg. 21, p. 97 (Genus Staurocephalus). Grube, A. E. °78. Fortsetzung der Mittheilungen iiber die Famile Eunicea. Jahres-Bericht der Schles. Gesell. fiir vaterl. Cultur., Bd. LVI. pp. 109-115. (Stauro- cephalide.) 14 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. McIntosh, W. C. °85. Report on Annelida. Challenger Reports, Zool., Vol. XII. Stauro- cephalus, pp: 231-235; Palolo viridis, pp. 257-261. Malaquin, A. *93. Recherches sur les Syllidiens, pp. 1-477, 18 Pl’s. L. Danel, Lille. Mead, A. °97. The Early Development of Marine Annelids. Journ. Morphology, Vol. XIII. pp. 227-326, Plates X.-XIX., Figs. in text. Verrill, A. E. °82. New England Annelida. Trans. Connecticut Acad. Sci. Vol. IV. pp. 285-324 e, Plates IIJ.-XII. Webster, H. E. °79. On the Annelida Chetopoda of the Virginian Coast. Trans. Albany Inst., Vol IX. pp. 202-269, 11 Plates (Staurocephalus). Whitmee, S. J. °75. On the Habits of Palola viridis. Proce. Zoél. Soe. London, pp. 496-502. Wilson, E. B. °92. The Cell-Lineage of Nereis. Journ. Morphology, Vol. VI. pp. 361-480, Plates XIII.-XX., Figs. in text. Wilson, E. B. "98. Considerations on Cell-Lineage and Ancestral Reminiscence. Annal. N. Y. Acad. Sci., Vol. XI. pp. 1-27, Figs. in text. Fig. Fig. Fig. Mayer. — Staurocephalus gregaricus. oo on PLATE 1. Staurocephalus gregaricus, nov. sp., natural size, swimming near the sur- face of the water before the rising of the Sun. The terminal segment has broken off, and the genital products are escaping through tle orifice. Staurocephalus gregaricus, natural size, showing the worm in the act of expelling its sexual products. The eggs or sperm escape into the water through the nephridial tubules, and also through rents and tears in the cuticula of the worm. This contraction usually occurs immedi- ately after the rising of the Sun. Side view of the head end of the worm; magnified. (m) mouth. Sete of the parapodia. (a) are most dorsal; (b) next; (d) next; and (ec) most ventral. See Figure 13, Plate 2. Section of an embryo in the 16-cell stage, magnified 100 diameters. Age 3 hours. Section of an embryo in the gastrula stage immediately before the closure of the blastopore. (bp) blastopore; (sge) segmentation cavity. Age 93 hours. Longitudinal dorso-ventral section of an embryo 34 days old, magnified 100 diameters. (an) place where the anus is destined to appear; (g/) head glands ; (m) place where the mouth is destined to break through; (oes) cesophagus. (sf) mid gut, or “stomach.” The egg-membrane persists as a larval cuticula. Longitudinal dorso-ventral section of a young worm 16 days old. (an) anus ; (ds, ds, ete.) dissepiments; (g/) head glands; (m) mouth; (n) ventral nerve-chain ; (oes) esophagus ; (st) cavity of mid gut. Longitudinal dorso-ventral section through the head region of a mature worm, showing tentacular cirrus and muscular pharynx. The intes- tine of the sexually mature worm is practically empty. (m) mouth. (n) ventral nerve-chain. _ MAYER-ATLANTIC PALOLO” 2 ee PLATE, 1. memes t cone B Meisel lith Bostes Mayer. — Staurocephalus gregaricus. PLATE 2. Fig 10. Dorsal view of Staurocephalus gregaricus, nov. sp., magnified 2 diameters. Showing the sexual segments contracted after the expulsion of the genital products. Fig. 11. Dorsal view of head, showing mouth opening. Magnified. Fig. 12. Ventral view of head. Magnified. Fig. 13. Side view of parapodium of the 40th segment from the head of the worm. (np) nephropore. Fig. 14. Larva one day old. Showing the green-colored gland cells between the eyes. Fig. 15. Larva 33 days old. (gl) head glands. Fig. 16. Larva 5} days old. Figs. 17. (a) and () sete of a larva 5} days old. es ATE. 2. Pr B Meisel lith Beste, A a = ‘ Ni y MAYER-ATLANTIC’ PALOLO” ’ met Fig. Fig. Fig. Fig. Fig. Fig. 5 Fig. Fig. Fig. Fig. Mayer. — Staurocephalus gregaricus, 18. 19. PLATE 3. Larva of Staurocephalus gregaricus, nov. sp., 10 days old. Dorsal view of a young worm i6 days old. The animal now ceases to swim through the water, but will readily burrow beneath the surface of sand upon the bottom of the aquarium. Length 0.8 mm. Ventral view of the head end of a young worm 16 days old, showing the “jaws” of the esophagus. Dorsal view of a worm 26 days old. Length 1.2 mm. Side view of a worm 26 days old, showing the “ jaws ” in the esophagus. (m) mouth. Side view of the posterior segment of a worm 26 days old, showing cirri. Dorsal view of a worm 34 days old, showing the beginnings of the mature coloration immediately back of the eyes. Length 1.6 mm. Seta from a worm 26 days old. Dorsal view of the dorsal “jaws” from the esophagus of a worm 26 days old. Side view of the dorsal “ jaws ” from the esophagus of a worm 26 days old. PLATE. 3. MAYER-ATLANTIC’ PALOLO:”” B Meisel lith Basten.” ee aT. ee — = _ : e - i a % -_ — @ 4 = y 77 i Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE. Vou. XXXVI. No. 2. om NORTH AMERICAN FRESH-WATER RHYNCHO- BDELLIDZ, AND THEIR PARASITES. By W. E. CastLe. Wirn Ereut PLATEs. CAMBRIDGE, MASS., U.S. A.: . PRINTED FOR THE MUSEUM. r Aveust, 1900. apa a No. 2.— Some North American Fresh-Water Rhynchobdellide, and their Parasites. By W. E. CAstLzE CONTENTS. PAGE I. Introduction. F 18 e. Nervous System . ipeeviethods: . . . « 18 4. Glossiphonia heteroclita L. Ill. Classification 20 a. Habitat, Form, Size, Key to Species 20 Color . +a cre IV. Description of Species 21 b. Rings, Somites, Eyes, 1. Glossiphonia stagnalis L. 21 Suckers, etc. ; a. Habitat, Form, Size, c. Reproductive Organs . Color . 5A d. Digestive Tract . b. Rings, Somites, Eyes eee e. Nervous System . ce. Dorsal Gland, Suckers 23 5. Glossiphonia elegans Ver- d. Reproductive Organs . 24 rill See A e. Digestive Tract . 26 a. Habitat, Size, Color f. Nephridia 27 b. Surface, Rings, Somites, g- Nervous System . 28 Eyes, Suckers . h, Metamerism : 28 c. Reproductive Organs . (4) Number of Somites 28 d. Digestive Tract . a. Structure of a Typ- e. Nephropores, Nervous ical Ganglion . 29 System £1 Oe 8. Fused Ganglia . 29 6. Glossiphonia parasitica (2) Somite Limits 3l SAV) Uae 2. Glossiphonia fusca, sp. nov. 34 a. Habitat, Form, me. a. Habitat, Form, Size, 6. Rings and Somites . Color . : 34 c. Eyes, Mouth, Oral b. Rings, Somites, Byes, Sucker Suckers . 386 d. Reproductive Organs . : e. Reproductive Organs . 36 e. Digestive Tract . d. Digestive Tract 37 J. Nephropores, Nervous e. Nervous System . 38 System 8. Glossiphonia eee sp. g. Papille, Coloration . nov. ; 39 (1) Var. plana . a. Habitat, Benny Size, (2) Var. rugosa Color . . . . 389} V. Mutual Relationships of the 6. Rings, Somites, Eyes, Species Described . Suckers d 40| VI. Parasites . c. Reproductive Organs : d. Digestive Tract . 41 | Bibliography 41 | Explanation of Plates 1 Contributions from the Zodlogical Laboratory of the Museum of Comparative Zodlogy at Harvard College, E. L. Mark, Director, No. 112. VOL, xxxvi. — No. 2. 18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. I. INTRODUCTION. In the fall of 1897 a small leech, which is very abundant in the ponds about Cambridge, Massachusetts, was selected as an object for study in the class in Microscopical Anatomy in Harvard University. This selec- tion brought under my observation a rather large number of leeches living or prepared in one of various ways, and gave occasion to the studies out of which this paper has grown. The kindness of friends has greatly aided me in obtaining material. In this connection my thanks are due to Mr. G. M. Allen, who sent me living leeches from the White Mountains in New Hampshire and also collected for me much valuable material in Massachusetts; to the Museum of Comparative Zodlogy for the privilege of studying its collection of leeches ; to Professor James G, Needham, who sent me collections made in New York and Illinois, and also loaned me for study the collection of leeches belonging to Lake Forest University ; to Dr. C. A. Kofoid, who obtained for me leeches from Havana, Illinois; to Mr. R. H. Johnson, for specimens collected in Lake Chautauqua, N.Y. ; and last but not least, to Professor E. L. Mark and Dr. Otto Zur Strassen, who collected and preserved for me indi- viduals of several European species. Professor Whitman, who has given so much attention to the study of leeches, several years ago (’91*) pointed out the inadequacy of all descriptions then existing of our North American species of “ Clepsine,” showing that the descriptions in question were based on characters alto- gether too superficial and unreliable. Whitman himself presented a model in his description of “Clepsine plana;” but as this has not been followed by any similar account of our other species, I have thought it worth while to record in this paper some observations of my own, to- gether with the views regarding the external morphology and relation- ships of our common species, to which studies, chiefly anatomical, have led me. II. METHODS. For the study of the general anatomy of a leech and particularly for the study of its external morphology, it is important to have both living animals and those which have been killed in a good state of extension. Of the former I have been fortunate enough to obtain an abundance ; in preparing the latter I have found very serviceable the method recom- CASTLE: NORTH AMERICAN RHYNCHOBDELLID&. 19 mended by Lee (94, p. 17) of stupefying with carbonated water. The animals are placed in a Stender dish and covered with water from a “soda siphon.” As soon as they are thoroughly stupefied, they should be quickly transferred to the killing fluid, which is best used warm, not boiling hot, but heated to about 70°C. A stay of from two to five minutes in the carbonated water usually suffices to stupefy the smaller species enough for successful fixation, and indeed is better than more prolonged treatment. For if the animal still possesses a slight degree of irritability, it will usually straighten out in the warm killing fluid and die in a better state of extension than it was in before. The large species require a much longer treatment with the carbonated water. The best reagent to use in killing animals for whole preparations is, in my experience, Perenyi’s fluid, which leaves the animal well extended and renders it clear and transparent. It has the property of removing pigment from the body, particularly the darker sorts of pigment. For instance, I have noticed that in killing the beautifully variegated Glossi- phonia parasitica with this fluid, the green and brown spots often dis- appear entirely ; while the yellow and orange spots remain conspicuous. This quality is sometimes an advantage, sometimes a disadvantage. If one wishes to preserve the color-pattern unimpaired, he would do well to use a fluid containing picric acid, which seems to have the property of fixing the pigment ; or, better still, use formaldehyde both as the killing and as the preserving fluid. Flemming’s fluid is perhaps, on the whole, the best fixing fluid to use in preparing sections; corrosive sublimate is also good; Perenyi’s fluid is for this purpose not to be recommended, except for the study of the gross anatomy of the central nervous system, which it makes very clear by bringing out nerves and fibre tracts in strong contrast to their con- nective-tissue sheaths. Iron hematoxylin is the best stain which I have tried for sections. For whole preparations, animals should be heavily stained with carmine and then pretty thoroughly decolorized. I find Mayer’s hydrochloric acid carmine (70% alcoholic) very convenient and serviceable, as it stains powerfully and there is no danger of maceration of tissues, however long the stain is allowed to act. Decolorizing is best done with alcohol pretty strongly acidulated, as greater contrasts are thus obtained. I use 1% hydrochloric acid in 70% 1 This method of stupefaction is also very useful in the study of the living animal, for the leech may be kept entirely motionless in the carbonated water within a live-box for hours, and then be revived by placing it again in fresh water. 20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. alcohol, allowing it to act until the specimens have a light rose color, then wash well in neutral alcohol (90%), clear in cedar oil, and mount in balsam. III. CLASSIFICATION. Leeches of the family Rhynchobdellidz may be distinguished from all others by the fact that they possess an exsertile proboscis (pr’d., Figure 1), with the aid of which they obtain their food, for they are entirely with- out jaws such as the medicinal leech possesses. Our common North American species of this family belong to the genus Glossiphonia John- son (16), better known to many by its synonym Clepsine Savigny (20). Leeches of this genus have usually a broad flat body, which, when the animal is disturbed, is rolled into a ball. Each somite con- sists typically of three distinct rings ; but the somites at the ends of the body always contain a smaller number of rings. These leeches are found in the shallow water of ponds and rivers underneath stones, sticks, or leaves, or adhering to the bodies of their hosts. The smaller species feed upon snails, crustacea, or other small fresh-water animals ; the larger species are known to feed upon turtles, to whose shells they are often found attached. They probably suck the blood of other aquatic animals also. The following key may aid in distinguishing the species to be described : — Key to Species. A. Crop diverticula a single pair (after a full meal the animal may have five more pairs, inconspicuous, and situated anterior to the prin- cipal pair) ; male and female genital pores separated by a single body ring ; rings without metameric markings in the living animal. 1. Eyes two, distinct ; a conspicuous yellowish brown chitinous spot on the neck dorsally . . . . . . G. stagnalis (p. 21) 2. Eyes two, inconspicuously pigmented or entirely without pig- ment ; no chitinous spot on the neck ; body extremely slender and transparent... . . . . . G. elongata (p. 39) B. Crop diverticula six pairs ; ‘sale anit female genital pores separated by a single body ring or else united. 3. Eyes two, the middle (sensory) ring of each somite marked throughout the greater part of the body by a transverse row of whitish spots... «© 6 s/s «06 a a) see CASTLE: NORTH AMERICAN RHYNCHOBDELLID#. 21 4, Eyes six, the first pair small and close together, the others farther apart; rings without metameric markings, or with dark pigment on the anterior ring of each somite. G. heteroclita (p. 42) C. Crop diverticula seven pairs ; male and female genital pores separated by two body rings. 5. Eyes six, distinct, in two parallel rows; a conspicuous longi- tudinal band of dark pigment on either side of the median plane dorsally, and a fainter one seeks ; inconspicuous papillz on the dorsal surface .. . oe al a "Golegasia (p. 46) 6. Eyes apparently a ails pair, far forward on the head and confluent ; back distinctly papillose. A large species, often found Mbuurees . « . « = «+ «= « - « « G. parasitica (p. 51) IV. DESCRIPTION OF SPECIES. 1, Glossiphonia stagnalis Liyyzvus (1758). Plate 1, Figs. 1, 3; Plate 2, Fig. 4; Plate 3, Figs. 7-10, 12; Plate 8, Fig. 34. Hirudo stagnalis Linneus (1758); H. bioculata O. F. Miiller (1774); Clepsine bioculata Savigny (’20); C. modesta Verrill (72); C. submodesta Nichol- son (’73). a. Hasitat, Form, Size, Conor. This species is found in Europe, the adjacent parts of Asia and Africa, and in North and South America. As one might expect in the case of so cosmo- politan a form, much has been written about it, but its external morphology has never been carefully and accurately analyzed, and published accounts of its internal anatomy contain a number of errors or omissions, some of which | hope to rectify. The general form of the body as seen in dorsal view, when partially extended, is shown in Figures 1 and 4. The body is broadest posterior to its middle and thence tapers gradually toward both ends. The head, which is only slightly wider than the neck, is evenly rounded in front (Figure 3) ; dorso-ventrally the body is very much flattened, especially when at rest. The animal is very active in its movements and can greatly elongate its body so as to become more than ten times as long as it is broad. The largest individuals measure as follows : — Length, fully extended, 20-25 mm, ; at rest, 8-10 mm. Width, fully extended, about 2 mm.; at rest, about 5 mm. Color, flesh-color or grayish. Small individuals are usually quite clear and transparent, but larger ones are apt to be more or less opaque. This opacity, 22 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. as well as the general grayish tint which the body often has, is due to the presence in varying proportions of two different sorts of pigment cells. Those of one kind, which might properly be called reserve-food cells, may be found in the deeper parts of the body of all well-nourished individuals. They are large rounded cells, with an excentrically placed nucleus, their cytoplasm being filled with rounded, highly refractive granules often nearly as large as the nucleus. By reflected light these granules appear of an orange-brown color. Osmic acid browns slightly, but does not blacken them. Corrosive-acetic or picro-nitric mixtures make their composite nature apparent. An outer shell of darker, brownish substance appears surrounding usually one, sometimes two or three perfectly clear spherical inclusions. Perenyi’s fluid, which is very strong in nitric acid, if allowed to act for about an hour, destroys almost every trace of the granules, the outer shell being the last part to disappear. Absolute alcohol acts in a similar way, but more slowly. Graf (99) has figured the granules accurately (see his Figures 87 and 102), but interprets their structure somewhat differently, regarding the clear portions as cavities; hence he speaks of the granules containing them as ring-shaped structures. I at first supposed the clear portion to be a central core unaffected by the killing fluid, but abandoned this idea when I discovered two or more of them in different parts of the same granule. It seems to me that the outer part of the granule, which possibly contains some fatty material, as the osmic acid test indicates, is laid down upon a central core of a different substance which dissolves much more readily in acid solutions. So much my preparations indicate, but do not prove conclusively. Further study should be given to these interesting structures, doubtless a reserve-food product, which reminds one of the structures found in the seed of the Castor-oil Bean (Ricinus). The second sort of pigment cell found in this species belongs to Graf’s (99) category of “excretophores.” They occupy a superficial position in, or just under, the epidermis, and are slender, thread-like, branched (structures) of a dark-brown color. They are especially abundant in animals which have been kept for some time in well-lighted aquaria. Graf believes that pigment cells of this sort become detached as leucocytes from the wall of the body cavity, take up excretory products in the deeper parts of the body, especially in the neighborhood of the blood vessels, and then by amcboid movements make their way to the surface of the body, there to disintegrate. b. Rives, Somires, Eyes. External rings, rounded and distinct; sixty-seven in number, counting two narrow rings at the posterior end of the body (64 and 66, Figure 4, Plate 2). Somites, thirty-four, as in all species of Glossiphonia. Somites VI.—XXIV., triannulate (Figure 4); all other somites show more or less abbreviation." 1 Throughout the descriptive part of this paper I shall speak of those somites which contain fewer than three distinct rings as “ abbreviated” or “reduced.” I a CASTLE: NORTH AMERICAN RHYNCHOBDELLIDZ. 23 Somites 1. and 11. are together represented by a single broad ring (Figures 3, 4), which, however, is sometimes subdivided by a shallow furrow (Figure 7, Plate 3). Somites 111. and Iv. consist each of a single ring, the latter forming the pos- _ terior boundary of the oral sucker (Figure 3, Plate 1; Figure 7, Plate 3). Somites xxv. and xxvi. consist each of two rings, a broad followed by a narrow one (63 and 64, 65 and 66, Figure 4, Plate 2; Figure 34, Plate 8). The narrow ring of somite xxv1., however, is often so completely fused with the broader ring which precedes it as to be scarcely distinguishable. Somite xxvii. consists of a single broad ring, crowded back to a position lateral and posterior to the anus (67, Figures 4, 34, and A). Somites XXVIII.-XXxXIV. are not represented by external rings; in the central nervous system, however, we shall find clear evidence of their separate existence, A further discussion of the metamerism will be deferred until the nervous system has been described. Eyes, two, large and distinct, lying in the anterior part of ring 3 and ex- tending forward into the posterior part of ring 2 (Figures 4, 7). c. DorsaL GLAND, SUCKERS. Dorsal Gland. — Between the twelfth and thirteenth rings (that is, between the anterior and middle rings of somite vi1r.) on the mid-dorsal surface of the arimal, is a structure (gl. d., Figures 4, 7) peculiar to this species, though accord- ing to Apathy (88+) its homologue is found: in some other species, either as a functional structure in the embryo, or as an inconspicuous rudiment in the adult. It consists of a rounded, wart-like, yellowish-brown, cuticular plate, often surrounded by a ring of substance similar but lighter in color, probably because less well hardened. These structures are secreted by a patch of high columnar epidermal cells, which in the embryo, according to Apathy, form a sort of byssus gland serving to attach the young to the under side of the mother before the suckers at the ends of the body become functional. In the adult the organ has no known function, thongh it forms a favorite place of attachment for a certain colonial protozodn of the genus Epistylis. do so, however, without feeling at all certain that the terms are strictly appli- cable in all cases or even in a majority of cases. I have elsewhere (Castle, 1900) expressed the opinion that the leech somite consisted primitively of a single ring. If this is so, it may well be that the somites commonly spoken of as abbreviated have really never attained the triannulate condition. (Moore, 1900, has expressed a similar view since this paragraph was written.) Nevertheless the term is a con- venient one to express deviation from the typical condition of the somite in the direction of a shortening of it. In this sense the term will be employed in this paper. 1 Budge (’49) likewise represents the eyes in the anterior part of ring 3. Apathy (’88*), however, counts the ocular ring the fifth, emphasizing subdivisions which can occasionally be seen in the most anterior rings. (Compare his Figures 4 and 10 with my Figures 3 and 7.) 24 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. The oral sucker (suc. or., Figure 7) lies on the ventral side of the head, within the limits of rings 1-3 (somites I.—Iv.). The mouth (or., Figure 7) opens anterior to the middle of the oral sucker as well as anterior to the eyes. The posterior sucker (act., Figures 1, 4), also ventral in position, is slightly longer than broad. Average dimensions for the largest individuals are :— length, 1.31 mm. ; width, 1.24 mm. d. REPRODUCTIVE ORGANS. The male genital pore (po. &, Figure 4) lies in a mid-ventral position between rings 24 and 25; that is, between the anterior and middle rings of somite x11. The female genital pore (po. 9, Figure 4), which is a broad transverse slit, lies just one ring behind the male pore, between rings 25 and 26, the middle and posterior rings of somite x11.) XII. XVIII. Testes (Figure 4, te.), six pairs, placed intersegmentally in somites =e The size and appearance of the testes vary considerably with the seasons. In the fall and early spring they are generally large and their outlines more or less irregular, for they adapt themselves to the spaces left them among the dorso-ventral muscles and other deep-lying organs. The testis wall is quite thick on its dorsal, ventral, and lateral aspects, but somewhat thinner on its median aspect. It is lined with a loose germinal epithelium of spindle-shaped cells, except at its dorso-median angle, where there is a small patch of ciliated epithelium continuous with that of the vas efferens. Male genital ducts. — The vas efferens is a short, delicate tube, which leads dorsad and cephalad to join a longitudinal duct similar in structure to itself and only slightly larger, the proximal or collecting part of the vas deferens (Figure 4, va. df.). Anterior to the first pair of testes, that is, about on the border between somites x1. and x1mI., the collecting portion of the vas deferens bends sharply toward the median plane of the body and passes between the strong dorso-ventral muscles, which, like a row of pillars, mark off on each side the 1 I am unable to find in any published account an explicit statement as to the position of the genital pores in this species. Budge (’49) figures the male pore in the posterior third of ring 25 and says, “ Gegen den 25. Ring findet sich die sehr feine mannliche Geschlechtséffnung.”’ He does not figure the female pore, but says (p. 100), “‘Ungefahr am 27. Leibesringe die aussere [female] Geschlechtsdff- nung liegt.” This would make the genital pores distant from each other about two rings, which, however, is incorrect. Ludwig (’86) incorrectly describes the position of the genital pores for the entire genus ‘“Clepsine” as follows (p. 781) ‘“‘minnliche Geschlechtséffnung zwischen dem 25. und 26., weibliche zwischen dem 27. und 28. Ringel.” This state- ment rests upon two erroneous assumptions, first, that the number of distinct rings is the same in the head region of all species, and, secondly, that the genital pores are always tworings apart. In only two of the six species described in this paper are the genital pores separated by two rings. CASTLE: NORTH AMERICAN RHYNCHOBDELLIDZ. 25 lateral limits of the median lacunar space. This space the vas deferens enters in company with the ducts of the salivary glands, which here pass inward to join the base of the proboscis (Figure 1, gi. sal.) Having reached the median lacuna, the vas deferens turns backward, running usually ventral and lateral to the diges- tive tube and parallel with the course of its collecting portion. In the median lacuna it winds about more or less, or may even cross into the opposite half of the body as a result of its being crowded for room either because of its own dis- tended condition or from the condition of other organs in its vicinity.. As it runs backward it widens into a spacious seminal vesicle (Figure 4, vs. sem.), and its epithelial lining ceases to be ciliated. The dimensions of the seminal vesicles vary with the amount of sperm stored in them, being capable apparently of great enlargement. Sometimes the vesicle runs back as far as the pair of long crop diverticula in somite x1x. (Figure 1), and is crowded out in the form of one or more loops between the testes (Figure 4); it may even find room for itself by crossing into the opposite half of the body. Ultimately it bends for- ward again and, narrowing, continues as the muscular and glandular ejaculatory duct (Figure 4, dt.e).). The ejaculatory duct, as it runs forward, passes out- side of the inner row of dorso-ventral muscles at about the point where the collecting portion of the vas deferens enters the median lacuna. It then runs forward into somite XI., where, turning sharply back again, it expands into a thick-walled “terminal horn,” which, uniting with the terminal horn of the other half of the body, opens to the outside by the mid-ventral male genital pore (po. gf, Figure 4). The special function of the ejaculatory duct and par- ticularly of its terminal horn, Whitman (’91) has shown to be the formation and extrusion of the spermatophore. In the early spring, as the water in the ponds begins to grow warmer, the seminal vesicles are seen to be gorged with sperm, and the formation of sper- matophores takes place rapidly. These the animals attach to one another’s backs. Whitman (’91) has shown that in the case of G. parasitica (‘‘ Clepsine plana”’) the contents of the spermatophore pass through the integument into the body cavity, and that impregnation probably occurs while the egg is still in the ovary. A similar process doubtless occurs in the case of G. stagnalis. _ After the period of active spermatophore formation has passed, — it ordinarily lasts but a few days or weeks, depending upon the rapidity with which the temperature of the water rises, — the vasa deferentia are seen to be greatly re- duced in size and the testes quite inconspicuous, though in the fall they were the most conspicuous organ in the entire body. The ovaries (Figure 4, oa.) are a pair of simple sacs extending back from the female genital pore in the median lacuna, usually ventral and lateral to the digestive tube. They are attached more or less loosely by mesenterial strands of connective tissue to those portions of the vasa deferentia which lie in the median lacuna. This connection, however, is so slight that when crowded for room an ovary may extend out in loops between the testes, or across into the opposite half of the body, just as the vasa deferentia do. The size of the ovaries depends upon the state of maturity of the contained ova. They are 26 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. largest in the early spring immediately before the eggs are laid, when they often extend the whole length of the genital region and are looped or folded, as are the seminal vesicles ; they are smallest immediately after the egg-laying. A mean between these two extreme conditions is shown in Figure 4. The time of egg-laying, as well as of spermatophore formation, depends upon the warming of the water in the spring. One can hasten both processes by bringing the animals for a few days into a heated room. Around Cambridge the eggs are laid mostly in the months of April and May. Small-sized indi- viduals, however, may come to maturity later in the season, even as late as September. The eggs are pink in color and about 0.3 mm. in diameter. They are attached to the under surface of the body in groups of two to eight eggs each. Each group is enclosed in a separate, delicate, transparent sac, which adheres to the under surface of the body. The sacs are arranged in two longitudinal rows close together, one on either side of the median plane of the body. The more posteriorly placed sacs usually contain more eggs than those farther forward. I have not observed the process of egg-laying, but believe that the eggs of a single sac are laid at about the same time, that they are then crowded back as far as possible under the body, and that there is poured over them a secretion from the clitellar glands which hardens into the delicate wali of the sac. After a period of rest, during which the body is closely applied to the group of eggs so that its sac becomes fastened to the body, another group of eggs is laid, and so on until all the mature eggs have been expelled from the ovary. The cli- tellar glands are deep-seated, unicellular epidermal glands opening on the ventral surface in the vicinity of the female genital pore. They can be demonstrated by methylen-blue staining. Animals which are kept in aquaria lay their eggs at night, and always com- plete the process in a single night, so that all the eggs borne by an individual are in about the same stage of development at one time. I think it probable that the egg sacs are arranged in the order laid, from behind forward, for in one of the most anterior sacs a single egg is occasionally found, but never in one of the more posterior sacs have I observed so small a number. The number of eggs laid by an individual depends upon its size. An animal thirteen mm. long (when fully extended) was found to have laid sixteen eggs; another twenty-six mm. long was found carrying forty-five eggs. The average number for nine individuals examined at one time was thirty-one. The usual number of egg sacs formed is six or eight; in one case examined it was ten. The average number of eggs found in a sac is about four ; for the most anterior pair of sacs it is three. e. Dicestive TRACT. The position of the mouth (or, Figures 3, 7), except when the body is much contracted, is anterior to the eyes, in the third somite (ring 2). It leads dorsally into the pharyngeal sac (sac. phy., Figure 7), which con- CASTLE: NORTH AMERICAN RHYNCHOBDELLID#. 27 tinues backward through the brain mass, ending in somite x11. (Figure 1). Within the pharyngeal sac lies the proboscis (pr’b., Figure 1), which, in a state of rest, usually extends from a point just behind the brain back into somite xur., where the ducts of the salivary glands enter its walls. These glands (gl. sal., Figure 1) are a conspicuous feature of a Glossiphonia differentially stained. ‘They are always unicellular, and represent the largest cells found in the body except certain nephridial cells and eggs approaching maturity. The salivary gland cells have a great avidity for stains. They number in this species thirty or more in each half of the body, and are found scattered through about three somites (x11.-x1v.). The largest gland cells are those most remote from the base of the proboscis. Each cell has a separate slender duct leading into the wall of the proboscis and opening into the lumen of that organ at some point along its length. A short slender esophagus (@., Figure 1), ordinarily lying entirely within somite xI1I., connects the base of the proboscis with the crop (@glo., Figure 1). This readily distensible part of the digestive tract extends over six somites (x1v.-xrx., Figure 1). Under ordinary circumstances it has but a singie pair of lateral diverticula ; these arise in somite xIx. and extend backward, usually ending in somite xxi. After a full meal, however, short lateral diverticula may sometimes be seen also in the five more anterior somites (XIV.-XVIII.), but this condition appears always to be a transient one. The stomach (ga., Figure 1) begins in somite xx. and ends in somite XXIII. ‘It bears four pairs of persistent lateral diverticula doubtless originally seg- mental in origin, but now crowded within the limits of about three somites. The first two pairs of stomach diverticula are directed forward, the last pair backward ; the third pair lies about at right angles to the long axis of the body. The terminal part of the digestive tract, the intestine (in., Figure 1), is a gradually narrowing tube; it includes one or two proximal chambers separated from the following part by constrictions. The anus is dorsal in position, as in all other leeches, and lies within or just behind somite xxvu. (Figure A, page 32; Figure 34, Plate 8). Comparison with other species, in which the reduction of somites is less extensive, shows that primitively the anus lay behind somite xxvit. jf. NepHRIDIA. The nephridia number at least sixteen pairs, possibly seventeen pairs. The nephropores (nph’po., Figure 4) lie on the ventral surface of the body, somewhat nearer the margin than the median plane, and almost exactly in the middle of their respective rings. The nephropores are always found in this genus on the middle ring of a somite. I have found them in sections of G. stagnalis in somites VIII.—-XXIV., with the single exception of somite x11. (ring 28). The strong development of the salivary glands in this region may account for the possible disappearance of the pair of nephridia which we should expect to find here. 28 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. g. NERVOUS SYSTEM. The central nervous system, as in other leeches, consists in the middle part of the body of a ventral ganglionic chain of twenty-one distinct ganglia meta- merically arranged and joined by paired connectives. Forming an extension of this ganglionic chain at either end of the body, one finds a nervous mass representing several primitively distinct ganglia more or less intimately fused together. In the central part of the body the ordinary position of the nerve ganglion is in the middle ring of its somite (Figure 4, somites XII—XVIII.). Toward either end of the body, however, there is a slight, but increasing, centripetal displacement of the ganglia, just as is frequently the case in the central nervous system of Arthropoda. This displacement may amount to as much as two-thirds of a somite, or in extreme cases an entire somite. Thus we see in Figure 4 that the ganglion of somite vu. lies in the first ring of somite vit, a displacement of two rings ; in somites vilI.—xI. the displace- ment is only a single ring. About the same amount of displacement occurs in somites XIX.-XXII.; in somites XXIII. and xxIv. it amounts to about two rings ; and in somites XXV.-XXVII. it is still greater. The positions in which the nerve ganglia are shown in Figure 4 are average ones carefully computed from the observed positions in five different individuals. The ganglia are very constant in position, the extreme variations usually amounting to only a fraction of the width of a ring. The structure and morphological value of the ganglionic masses at the two ends of the body is a subject closely connected with the general question of the metamerism of the body. h. MeTAMERISM. (1) Number of Somites. A number of investigators have discussed the question of how many somites are found in the body of a leech, and have reached conclusions varying accord- ing as they placed emphasis on one or another of the following criteria: (1) The number of external rings ; (2) color markings of rings, or the reeur- rence of peculiar papille on certain rings of each somite; (3) metameric sense organs; (4) the number of ganglia in the central nervous system as determined (a) by a count of the nerve capsules, typically six to a ganglion, or (0) by ascertaining the number and peripheral distribution of the nerves arising from the ganglia. Whitman (’92), making use principally of the criteria named under 3 and 4, was the first to obtain an entirely satisfactory answer to the question. He has shown that in the central nervous system of ‘* Clepsine hollensis” (which is closely related to G. parasitica) there are present thirty-four ganglia, each giving off paired nerves. Six of these ganglia are found in the anterior ganglionic mass which encircles the pharyngeal sac; seven are found in the posterior CASTLE: NORTH AMERICAN RHYNCHOBDELLID. 29 ganglionic mass which lies in the posterior sucker and supplies it with nerves ; these, added to the twenty-one distinct ganglia found in the central part of the body, bring the total up to thirty-four. An examination of the sense organs connected with these ganglia, and situated typically on the middle ring (first, Whitman) of each somite, yields corroborative evidence that the number of somites represented in the body is thirty-four. Bristol (99) subsequently made a similar study of the metamerism of Nephelis lateralis, his conclusions being for the Gnathobdellide entirely in harmony with those of Whitman for the Rhynchobdellide. Oka (94), however, has cast doubt upon the general applicability of Whit- man’s determination, based as it was on the metamerism of a single species of Glossiphonia, by stating that in the several European species which he has studied (G. stagnalis, G. complanata, G. concolor, G. heteroclita, G. papillosa, G. marginata, and G. tessellata) he finds evidence of only five (not of six) fused ganglia in the brain. Moreover, in recent systematic papers, such as those of Blanchard (94) and Moore (’99), we find the body of the leech still analyzed and described as consisting of twenty-six preanal somites, instead of twenty- seven, the number found in that portion of the body by Whitman ('92) and Bristol (99), and still earlier, though on less satisfactory evidence, by Apathy (88). Accordingly, I have thought it worth while to examine into this matter rather carefully in the case of the species studied by me. I may say at once that my results, in the case of all six species studied, are in complete accord with those of Whitman (’92), so far as the number of meta- meres is concerned. In determining the limits of the somite, I have arrived at conclusions differing from those of my predecessors, as will presently appear (p. 31 ff.). a. Structure of a Typical Ganglion.— A typical ganglion from the middle of the body has its ganglion cells arranged in six groups enclosed in capsules of connective tissue. Four of these capsules are lateral in position, two on each side of the ganglion ; the other two occupy a mid ventral position, one in the anterior, the other in the posterior part of the ganglion. (See the ganglion of somite XXvI. in Figure 9, Plate 3.) Three nerves are given off close together from either side of the ganglion, and are distributed to the three successive rings of one and the same somite, as I have elsewhere (Castle, 1900) pointed out. If, then, we can determine exactly how many such ganglia are present in the central nervous system of a leech, we shall be in a position to say how many somites enter into the composition of its body. In the middle part of the body, as already stated, twenty-one distinct ganglia of the sort just described can easily be recognized. To determine how many are present toward either end of the body, where more or less fusion of ganglia has taken place, is a matter of more difficulty. B. Fused Ganglia. — Figure 9 (Plate 8) shows a dorsal view of the poste- rior part of the central nervous system of G. stagnalis, obtained by reconstruc- tion from a series of frontal sections. The last two distinct ganglia, those of 30 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. somites XXVI. and XXVII., are shown, followed by the nerve mass of the poste- rior sucker, made up of seven fused ganglia. In it seven pairs of lateral capsules appear on either side, a segmental nerve root being closely connected with each pair (xXXvuI.-xxxIv.). The more posterior of the lateral capsules has in the case of each pair been displaced outward and downward (ventrad) and been reduced in size. The position of the seven pairs of ventral capsules is indicated by dotted outlines, the numeral denoting the somite to which each capsule belongs. In the first and last of the fused ganglia of this region, the ventral capsules occupy their typical tandem position (as in ganglion 26); in the case of the intervening ganglia (29-33), we find a more or less complete displacement of the ventral capsules to a side-by-side position. A similar dis- placement occurs in ganglion 27, which lies close back against the septum which divides the lacunar space of the posterior sucker from that in which the more anterior portions of the central nervous system lie. The same mechanical cause, crowding in an antero-posterior direction, explains both phenomena of displacement. The evidence presented in Figure 9 leaves no room for doubt that seven primitive ganglia are found in the nerve mass of the posterior sucker in this species. Determination of the number of ganglia represented in the brain mass is not quite so easy, but the evidence is likewise convincing. The brain (6., Figures 4, 7) forms a ring of nervous substance situated commonly in the last ring of somite vi. and the first two rings of somite vu. It surrounds the thin-walled pharyngeal sac (sac. phy., Figure 1), there being in leeches no recognizable separation into supra- and sub-cesophageal ganglia. A lateral yiew of the brain and the metameric nerves given -off from it is shown in Figure 8; a view of its dorsal surface in Figure 12. Figure 10 shows the arrangement of the capsules on its ventral surface. An examination of Figures 8 and 10 shows that the capsules (6, 6) of the last brain ganglion have quite their typical arrangement. A triple segmental nerve (v1., Figure 8) emerges from under a pair of lateral capsules, while below a pair of ventral capsules are arranged in the usual tandem order (6, 6, Figures 8, 10).? Ganglia 3-5 likewise present no special difficulties, their lateral capsules being present in pairs with nerve roots attached (3,3 4, 4; 5, 5, Figure 8). 1 IT have been unable to determine to what extent in the reduced somites at the two ends of the body the original triple nature of the segmental nerves per- sists. The nerve of the last brain ganglion is certainly triple (v1., Figure 8), as we should expect from the fact that somite vi. consists of three distinct rings (Figures 4,7). Most of the nerves ‘anterior to this one, perhaps all, are either double or triple, but as I have been unable to determine accurately which con- dition exists in some of them, I represent the nerve as undivided in the case of the first five somites (Figure 8). Fora like reason I follow a similar course in repre- senting the segmental nerves of the posterior ganglionic mass (Figure 9). I think that all of these nerves are made up of at least two distinct bundles of fibres ; whether the small third nerve is also present as a distinct element in any or all of them, I am unable at present to say. oo — aineei tes pel CASTLE: NORTH AMERICAN RHYNCHOBDELLID#. 31 Their ventral capsules show the following modification in arrangement ; they have been displaced from the typical tandem position to a side-by-side position (Figures 8, 10; compare Figure 9, somites xxIx.—xxx1II.). The lateral capsules of ganglion No. 2 are found dorsal to the pharyngeal sac (2, 2, Figures 8, 12). They seem to have been displaced backward to a po- sition somewhat posterior to the lateral capsules of ganglion No. 3 by a migration in that direction of the supra-cesophageal commissure (Figure 8 ; compare Figures 11, 21). The commissure in this species is normally thrust back of the position in which it is shown in Figure 8, so that it lies about over the lateral capsules of ganglion No. 5. The animal whose brain is represented in Figure 8 was curved ventrad so that the commissure was thrust forward of its usual position and the row of lateral capsules below it was straightened out a little. The position of the ventral capsules of ganglion No. 2 is shown in Figures 8 and 10; the nerve root (1I., Figure 8) arises at the anterior end of the brain just ventral to nerve root I. The ganglionic capsules of neuromere No. 1 ail lie dorsal to the pharyngeal sac and anterior to the supra-cesophageal commissure (Figures 8,12). I be- lieve that the most anterior and ventral of these (1v., Figures 8, 12), which lies closely attached to nerve root I. in each half of the body, is homologous with a ventral capsule of one of the succeeding ganglia. Capsule Iv. extends out lateral to, sometimes even ventral to, nerve root I., so that its end may appear in sections between nerve roots I. and II. Oka (94) states that he finds in the brain of “Clepsine” (Glossiphonia) always thirty nerve capsules, and he accordingly regards it as equivalent to jiwe fused ganglia and no more. Since G. stagnalis was one of ‘the species studied by him, I am unable to understand how he can have reached such a conclusion, unless he has overlooked altogether the capsules of somite 1. which lie anterior to the supra-cesophageal commissure. Both the number and arrangement of the nerve capsules, and the number and \ position of the nerve roots, show clearly that in G. stagnalis stx fused ganglia are represented wn the brain, and that in the entire body THIRTY-FOUR somites ar represented. (2) Somite Limits. It remains to explain the grounds on which the limits of the somites have been placed by me as indicated in Figure 4. Whitman (’85) pointed out many years ago that a certain ring (the first, according to his account) of each typical somite in the body of a leech is more richly supplied with sensory organs (“sensillz ”) than any other ring of the somite. In many species of Glossiphonia special color markings or papille are also found on the sensory ring. Color markings, however, are wanting in G. stagnalis, and the sen- sill are not sufficiently conspicuous in the living animal to make identification of the sensory rings at all certain. But a carmine stain of the proper intensity renders identification of the sensory rings quite easy by giving them, especially along the margins of the body, a somewhat darker color. Observing this fact, VOL, XXXVI. — NO. 2. 2 32 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. I was first enabled to determine as sensory the rings indicated by Arabic numerals in the right half of Figure 4 ; further study revealed the presence of marginal sensille in the positions indicated in Figure 3. The metamerically repeated sensory annuli were thus positively identified throughout the greater part of the body. It remained merely to mark off the somite limits between successive sensory annuli. This I at first did after the usage of Whitman (’85, ’92) and practically all others since the time of Gra- tiolet (62), considering the sensory ring as occurring at the anterior end of ats somite. I found, however, that a consistent following of this practice would, toward either end of the body, place the somite limits in the middle of a ring instead of between rings, the position in which somite boundaries fall in other regions of the body, See Figure A, xxv’., xxvr’,, etc. Figure A.— G. stagnalis. Dorsal view of posterior part of body, showing mar- ginal sensilla. Somite limits are indicated correctly at the right of the figure (xxiv. to xxvit.); at the left of the figure (xx111’. to xxvm’.) they are shown as they have been commonly but incorrectly placed. This led me to inquire whether the sensory ring really is the anterior ring of its somite. The results of this inquiry have been published elsewhere (Castle, 1900), so that only one or two of the more important conclusions need be restated here. One of these, already suggested in part on page 29, is the following : — Somite limits coincide with neuromeric limits ; consequently in Glossiphonia the sensory ring is the middle, not the anterior ring of the somite. CASTLE: NORTH AMERICAN RHYNCHOBDELLIDA. 33 This point being established, the somite limits must be marked off, in the regions where unabbreviated somites occur, as in Figure 4, vI.-xxtv. I have further shown, in the publication already cited, that in Glossiphonia somite abbreviation! is accomplished by a series of steps which follow one another in regular sequence. First, a union takes place between the sensory ring and the ring which precedes it; secondly, the ring which follows the sen- sory ring is reduced in size; finally, it too fuses with the sensory ring, the entire somite being then represented by a single external ring. If, as is not improbable, some of the “ abbreviated ’”’ somites are really in arrested stages of development from the one-ringed to the three-ringed con- dition (as suggested in the case of Microbdella by Moore, 1900), the order of the three steps enumerated should be reversed, in their case, and described in the following terms: (1) A distinct narrow ring is separated off at the pos- terior end of the unidnnulate somite; (2) this newly formed posterior ring grows in width ; (3) another new ring is separated off at the anterior end of the somite. This produces a three-ringed somite, all three rings ultimately attaining an equal width. For convenience in description, however, the pro- cess will be uniformly treated as one of abbreviation, as explained on page 22, footnote. The amount of “abbreviation,” as is well known, becomes greater toward either end of the hody. Bearing in mind these principles, we find that the least affected of the abbreviated somites of G. stagnalis are those which stand nearest to the un- abbreviated somites, namely, v. (Figure 3) at the anterior end of the body, and xxv. (Figure A) at the posterior end. In the case of each of these, the anterior and sensory rings of the somite are united into a single broad ring. But in the case of somite v. we find the union occasionally incomplete, as in- dicated by the notch (less clearly than it should be) at the upper margin of Figure 3, ring 4. Somite xxvi. (Figure A; Figure 34, Plate 8) is usually found in the same condition of abbreviation as the somites just described. Occasionally, however, its posterior ring is narrower or less distinct than that of somite xxv. In somites 111. and rv. the process of abbreviation to a single broad ring is practically complete, although the narrow posterior ring is in favorable pre- parations still recognizable as a distinct element separated from the rest of the somite by a shallow transverse furrow (Figure 7, III, Iv.). Somites I., 11., and xxvit. have each been reduced to a single ring; in addition a fusion (sometimes incomplete) has taken place between somites I. and 11., so that they are together represented by the broad ring, 1 (Figure 7). Somites XXVIII.-XXXIV. are not represented by annuli on the surface of the body ; they form collectively the posterior sucker. 1 As to the sense in which this term is used, see p. 22, footnote. 34 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 2. Glossiphonia fusca sp. nov. Plate 4. a. Hapitat, Form, Siz, Conor. This species is rather closely related structurally to G. stagnalis, with which I have found it associated in the vicinity of Cambridge, Mass., and Trenton, New Jersey. It is of about the same size as G. stagnalis, but is broader in proportion to its length (Figure 13, Plate 4). In its movements it is some- what more sluggish than that species and does not stretch itself to so great a length. Length of largest individuals, fully extended, 20 IMm.; at rest, 9 mm. Greatest width, when fully extended, 2.56 mm.; at rest, 4 mm. Color, a coffee-brown above, somewhat lighter below. The general brown coloration is due to the presence in the superficial layers of the body of slender, branched, thread-like pigment cells bearing numerous knot-Jike swellings and filled with a dark-brown pigment. Such pigment cells are clearly homologous with the pigment cells found in a superficial position in the body of G. stag- nalis, — Graf’s ‘*‘ excretophores.” They are much more abundant on the dorsal than on the ventral surface. On the former they appear in greatest numbers in a median dark band about as wide as two or three body rings; but they are entirely wanting anterior to the eyes and in the following regions, which therefore appear as clear, transparent areas : — 1. A transverse row of circular spots found on the sensory ring of each somite. These spots are about the width of a ring in diameter. Their maxi- mum number is seven, but it is a rare occurrence to find all seven present in a single somite. Each spot occupies a definite position on its ring, so that those of successive somites form seven longitudinal rows, three in each half of the body and one median in position. The paired rows may be designated as marginal, intermediate, and paramedian, for they occupy positions which cor- respond closely with those of the rows of dorsal papille so designated in the case of G. parasitica (Plate 2, Figure 6). : The paramedian rows of clear spots are more constant in occurrence than any of the others ; they can usually be found on somites v.-xxvi. The intermediate and marginal rows usually begin about in the region of the genital pores and continue with increasing distinctness back to the anus. The median row is less well developed than any of the others, It is represented by an oc- casional clear spot in the region posterior to the genital pores and anterior to somite XXII. 2. In the region of somites xx11.-xxvi., the median row of clear spots is suddenly replaced by a continuous clear band about as wide as one of the spots. Along the margins of this clear band, the pigment is unusually abundant, which fact adds by contrast to the conspicuousness of the median band. 3. The margin of the posterior sucker, where it projects beyond the outline of the body as seen in dorsal view, usually bears eight or ten triangular or CASTLE: NORTH AMERICAN RHYNCHOBDELLIDZ. 35 rounded clear spots of approximately the same form and position as the yellow pigment spots found on the posterior sucker of G. parasitica (see stippling in Figure 6). 4, The sensory ring of each of the somites in the neck region — somite v. and a few of the following — is occasionally distinguished by an uninterrupted, but narrow, clear band, which runs entirely across it from one side of the body to the other, occupying about its middle third. The conspicuousness of the unpigmented areas just described, except that mentioned under (4), is increased by the presence in tle centre of each of a group of peculiar reserve-food cells, which lie quite near the surface of the body. The ordinary reserve-food cells of this species agree in practically every par- ticular of structure and distribution with those of G. stagnalis, They are large rounded cells, sometimes attaining a diameter of eighty mikra or more. The granules within their cytoplasm attain a diameter of six or seven mikra. The color of these cells by reflected light is a pale orange; by transmitted light, they are semi-transparent, of a leaden gray color. They are distributed ir- regularly through the middle and posterior portions of the body, being situated in its deeper parts. The special form of reserve-food cell, which is found in the segmental clear spots already described, differs in respect both to size and to color from the ordi- nary reserve-food cell. It is considerably smaller, — forty to fifty mikra being the maximum diameter observed, — and its contained granules are likewise smaller, though more numerous. Its color by reflected light is a bright lemon yellow ; by transmitted light it is brown. Finally this variety of reserve-food cell is invariably situated quite near the surface of the body. The appearance of a group of these cells as seen under a moderately high power of the micro- scope is shown imperfectly in Figure 17 (Plate 4). The ventral surface of the body is pigmented in very much the same fashion as the dorsal, but less heavily. There is, however, this difference in the dis- tribution of the superficial brown pigment : on the ventral surface a pair of narrow, paramedian, pigmented lines can be recognized, one in each half of the body, in about the position of those found both dorsally and ventrally in G. elegans (Figure 30, Plate 7). On the dorsal surface, on the other hand, the most heavily pigmented region is a broad median band (p. 34). Segmental clear spots are found on the sensory rings on the ventral surface also, and these are arranged in paramedian, intermediate, or marginal rows; but the spots are much less conspicuous than on the dorsal surface, and the lemon-yellow reserve-food cells are less often found in their centres. Comparing the coloration of this species with that of G. stagnalis, we may say that the histological elements which produce the coloration are very similar in the two, but the distribution of these elements is such as to produce in G. fusca a distinct color pattern (longitudinal striations and segmental clear spots), a feature entirely wanting in G. stagnalis. 36 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 6b. Rines, Somires, Eyes, SucKERS. External rings, not quite so distinct as in G. stagnalis ; skin, slightly rougher owing to the stronger development of Bayer’s (’96) sense organs. Number of preanal rings, seventy (Figure 13, Plate 4). Somites V.-XXIv. are triannulate, but the two anterior rings of v. are united ventrally (Figure 15). Somites 1. and 1. are included in a single broad ring, which, just as in G. stagnalis, is sometimes subdivided by a shallow transverse furrow (Figure 14) marking the boundary between the two incompletely fused somites. Somites II., Iv., Xxv. and xxvi. (Figures 13-16) are biannulate. In each case the broader, anterior ring bears the sensille and corresponds to rings 1 and 2 of triannulate somites (compare somites Iv. and v. of Figure 15). Somite xxvil. is a single broad ring (70, Figure 13) which lies just anterior to the anus, not crowded back of it, as in stagnalis (Figure 34, Plate 8). The principal differences in somite composition between fusca and stagnalis occur in the head region, in somites 111.-vy. These somites are less abbreviated (or more fully elaborated) in fusca than in stagnalis, hence the greater number of preanal rings in the former (seventy) as compared with the latter (sixty- seven). Eyes, two, large and distinct, situated in rings 3 and 4 (Figures 14-16). The sensory elements of each eye, as in G. stagnalis, are contained in a pig- ment cup which is open only on its anterior, lateral surface, where the nerve fibres make their exit (Figures 14, 16). Oral sucker, as in all species of Glossiphonia, included within the first four somites (Figures 14, 15). Posterior sucker of about the same dimensions as in G. stagnalis, slightly longer than broad. c. REPRODUCTIVE ORGANS. Male genital pore (po. @, Figure 13), between the first and second rings of somite x1I. (rings 27 and 28). Female genital pore (po. 9, Figure 13), between the second and third rings of somite x1. (rings 28 and 29). Testes (te., Figure 13), six pairs situated intersegmentally in somites XII. XVIII. “iy... 20x) The ovaries have the usual form and position of these structures in other species, being found ventrally in the median lacuna. Eggs are laid a month or six weeks later than by G. stagnalis (June 12, 1898, Cambridge, Mass.). In color they resemble those of G. stagnalis closely, being of a light pink or flesh color. As in G. stagnalis, the eggs are attached to the under side of the body posterior to the genital pores, within a number of delicate sacs arranged in two parallel rows, close together, one on each side of the median plane. The number of sacs is most often six, but a seventh sac CASTLE: NORTH AMERICAN RHYNCHOBDELLIDA. 37 was observed in one case. The number of eggs in a sac, as well as the total number of eggs laid by an individual, is greater in this species than in G. stag- nalis. The following figures will indicate the number of eggs borne by four good-sized individuals, which laid eggs in the laboratory in June, 1898. The vertical line represents the median plane of the body; the positions of the numerals show how the sacs were placed with reference to one another and to the median plane of the body ; the numerals themselves indicate how many eggs were in each sac. Anterior is toward the top of the page, and the right side of the body toward the left of the page, the animals having been observed in ventral view. Inpivipvuat I. InpivipuaL II. a. a. 16 | 17 11 | 6 wm i6'| 20 2 ae 13 2. 22 | 13 18 | 14 p- ps Total 54 +50 — 104 44 4+33=77 Inpivipvat III. InpivipuaL IV. a. a. 5 2| 13 16 | 14 “P ‘18 | 20 Lp T. 21 | 19 lp 19 | 18 17 | 13 p- p- 39 + 51 = 90 54 + 51 = 105 Average number of eggs in a sac in above cases, 15 (as against 4 in G. stag- nalis); average number of eggs borne by an individual, 94 (as against about 30 in the case of G. stagnalis). It will be noticed that one of the anterior sacs often contains a relatively small number of eggs (as noticed in the case of G. stagnalis also), suggesting that it served to finish off the egg-laying, the sacs being arranged in the order in which they were formed, from behind forward. d. Digestive Tract. The mouth is situated anterior to the eyes, well forward in the anterior half of the oral sucker (Figures 14, 15). From here the thin-walled pharyngeal sae (sac. phy., Figure 13) leads back to the base of the proboscis in somite XIT., just behind the male genital pore. When the animal is at rest the proboscis (pr’b., Figure 13) usually extends through the four somites between the brain and 38 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. the male genital pore (VIII.-XI.) into somite x1I., where it receives the ducts of the salivary glands, a bundle from either side of the body. The salivary glands themselves are very large in this species and are dis- tributed in the marginal part of the body through somites XxI.—xvII., or, in exceptional cases, even a somite farther in one direction or the other.) The short esophagus (@., Figure 13) extends from the base of the proboscis through somite x11. to the beginning of the crop in somite xIv. The crop (i'glv., Figure 13) extends over the six somites xIv.-xIXx., giving off in the middle of each a pair of conspicuous lateral diverticula. These are always evident whether the crop contains food or not, a condition very different from that which exists in G. stagnalis. The last pair of crop diverticula (those of somite XIx.) are very long but simple, as in G. stagnalis, without secondary lateral diverticula. They extend back over the entire stomach region, usually ending in somite XXIII. The stomach (ga., Figure 13), which is separated from the crop by a valve- like constriction, bears four pairs of lateral diverticula doubtless originally metameric in arrangement, but now arising within the limits of somites XX.—XXII. The intestine (in., Figure 13) leads from the stomach back to the anus, which is situated dorsally just behind somite xXxvit., as in other species of Glossi- phonia. The intestine includes anteriorly two rather spacious chambers, the first of which bears a pair of small ear-like diverticula from its anterior lateral borders. Behind these chambers comes a simple tubular part terminating at the anus. To sum up, the particulars in which the digestive tract of G. fusca differs con- spicuously from that of G. stagnalis are (1) the shorter proboscis and larger cesophagus; (2) the larger salivary glands, distributed through a greater num- ber of somites; (3) the persistent character of the first five pairs of crop diver- ticula; (4) the distinctly chambered condition of the intestine, and the pair of diverticula borne by its first chamber. Nephropores are found on the sensory ring of each of the somites VIII.—XXIV., with the possible exception of x11I., where, as in stagnalis, the nephridia are much reduced, if not wholly wanting,—a fact accounted for by the strong development of the salivary glands and genital ducts in that region. The nephropore lies usually a little anterior to the middle of the ring on which it is found. e. Nervous System. A ventral view of the brain is shown in Figure 18, a dorsal view of that part of it which lies above the pharyngeal sac is shown in Figure 16, the position of the ventral part being indicated by a dotted line ; the outline of the brain 1 The animal shown in Figure 13 was a small one, and the salivary-gland cells are proportionally a little larger than they would be in the average, full-grown animal. ee a ee ee CASTLE: NORTH AMERICAN RHYNCHOBDELLIDA, 39 as seen in a lateral view is shown in Figure 14, cb. It lies for the most part in somites v1, and vit. ‘This is about a somite posterior to the usual position of the brain in G. stagnalis (Figures 4, 7). The number of fused ganglia represented in the brain is, as in G. stagnalis, six, and the nerve capsules have the same general arrangement as in that species. The ventral capsules of neuromeres 11.-v. are placed side by side, while those of neuromere vi. lie one behind the other (Figure 18 ; compare Figure 10, Plate 3). The six capsules of neuromere I. are situated well dorsal, as in G. stagnalis, and the supra-cesophageal connective is pushed back nearly over the middle of the entire brain mass (Figures 14, 16). The lateral cap- sules of neuromere Il. are shown in the dorsal view (Figure 16) ; those of neu- romeres III.—VI., in the ventral view (Figure 18). In Figure 14, which represents a parasagittal section, is shown the position of the paramedian sensille of the head somites, certain of which also appear in Figure 15. These indicate clearly the sensory rings of the somites in that region, and so aid in the determination of the external limits of the somites, The eye is clearly derived from one of the segmental organs of somite 11. (ring 2), as the position of its nerve indicates. This view is confirmed by a comparison with the conditions existing in G. heteroclita and G. elegans. 3. Glossiphonia elongata sp. nov. Plate 6. a. Hasitat, Form, Size, Conor. This leech first came to my notice in September, 1898. While collecting G. stagnalis from Spy Pond, near Cambridge, I found three or four leeches which, although of about the same size as stagnalis and occurring in similar situations, at once attracted my attention because of their more slender bodies and the peculiarities of their movements. These animals were carefully pre- served, and diligent search was made the following spring for more. This search, however, was fruitless; but in September, 1899, I was fortunate enough to find quite a number of individuals in a pool near Fresh Pond, Cambridge, some of which I have since kept alive in aquaria for several months. The body is less flattened dorso-ventrally in this species than in any other Glossiphonia known to me, being sub-cylindrical in cross-section. It is ex- tremely slender, even when contracted, and both head and acetabulum are small (Figure 27, Plate 6), This species does not roll itself into a ball, as other species do, when disturbed. Instead, it writhes about or twists itself into knots like an earthworm. In aquaria it moves little from place to place, but, attached by its weak posterior sucker, extends its snake-like body searching hither and thither as for a place of concealment, or, losing its attachment, seems unable to regain it and writhes helplessly like an earthworm on a smooth surface. The largest individuals which I have examined measure as follows : — 40 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Length, fully extended, 25 mm. ; partially contracted (as in Figure 27), about 10 mm. Width, fully extended, less than 1 mm.; partially contracted (as in Figure 27), about 1.5 mm. Color. — The anterior and marginal parts of the body are very clear and transparent. The rest of the body is usually of a pale yellowish-white color when the animals are first collected, but changes to a rusty yellow or pale orange color if they are kept in well-lighted aquaria for a few days. The color is due to the presence in the deeper parts of the body of rounded reserve-food cells, similar to those described as occurring in G. stagnalis. Apparently the nature of the granules in the reserve-food cells changes under the influence of daylight, so that by reflected light they appear pale orange instead of yellowish- white, the color which they have when first collected. Superficial pigment cells of the branched type, described as occurring in G. stagnalis and other species, appear to be entirely wanting in G. elongata. Fat cells occur in abundance in the deeper parts of the body, the contained oil drops being perfectly clear and transparent, as in G. stagnalis and G. fusca. 6. Rines, Somires, Eyes, SucKErs. The skin is very smooth and entirely free from papille. External rings, broad and smooth, usually indistinct in the head region (somites I.-Iv., Figure 23). Number of rings, 62 between oral sucker and anus (somites v.—XXVII.). Notwithstanding the indistinctness of the rings in the head region, favorable preparations, like that represented in Figure B, show that the composition of somites I.—Iv. is practically the same in this species as in G. heteroclita (Plate 5) and G. fusca (Plate 4). Somites 1. and II. are uniannulate; somites II. and Iv. biannulate, the anterior rings being broader and corresponding to rings 1 and 2 of a typical somite taken together. Somite vy. is likewise biannulate in this species, just as in G. stagnalis (Figure B; compare Plate 1, Figure 3); in all the other species with which this paper deals, somite v- is triannulate. Somites v1.-xx1v. (Figure 27) are triannulate, as in all other known species of this genus. Somites XxXv.-XXvII. are reduced each to a single ring, a con- dition found in the other species described only in the case of somite XXVII,, somite xxv. being always biannulate, and somite XXvI. usually so. Eyes, two, situated about as in G. stagnalis, just posterior to the mouth, between somites 11. and Iv. (Figure 23). The eyes are separated from each other by a considerable space, as in G. stagnalis (Plate 2, Figure 4) and G. fusca (Plate 4, Figure 16). The pigment associated with them is usually small in amount ; often it is wanting altogether. The oral sucker, as in the other species described, lies within the limits of somites 1.-1v. The mouth lies about in its centre (Figure 23, Plate 6; Figure B). The posterior sucker (act., Figures 24, 27) is extremely small and weak. In CASTLE: NORTH AMERICAN RHYNCHOBDELLID. 41 position it may be described as terminal rather than ventral (the position which it occupies in other species). c. REPRODUCTIVE ORGANS. The genital pores have the same position as in G. stagnalis and G. fusca; the male (po. g, Figure 27), between the first and second rings of somite xu., the female (po. 2, Figure 27), between the second and third rings of the same somite. Ficure B.— G. elongata. Ventral view of head end, showing annulation of head somites and position of marginal sensillz. Testes (te., Figure 27), six pairs placed intersegmentally in somites 0m XVII. XIV. XIX, The ovaries (oa., Figure 27) have the typical structure and position which they possess in other species (see p. 25). The eggs and egg-laying of this species I have not observed. » as regularly in the genus. . d. Digestive TRAct. The mouth (or., Figure 23, Figure B) opens about in the middle of the oral sucker. The proboscis (pr’b., Figure 27) commonly extends over about four somites (v111.-x1.). The salivary glands (gl. sal.) are found chiefly in so- mite x11., though a few may lie in the adjacent somites, xI. and x1m. About thirty good-sized gland cells are found in either half of the body. In size, number, and position the salivary glands of this species resemble those of G. 42 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. stagnalis more closely than those of any other species (compare Figures 1 and 27). The crop (2’glv., Figure 27), as in G. stagnalis, bears a single pair of diver- ticula, which arise in the middle of somite xrx.; but the diverticula are shorter in this species than in stagnalis, ending usually in somite xx. (com- pare Figures 1 and 27). The stomach (ga.), as in all species of Glossiphonia, bears four pairs of lateral diverticula. They arise within the three somites xx.-xxuI. All are directed slightly forward. The intestine (in.) is a simple tube not constricted into distinct chambers proximally as in most species. The anus (an., Figure 24) lies just behind somite xxv. In the structure of its digestive tract, as well as in the composition of its somites, this species shows a more reduced, simpler condition than is found in any other species known to me, stagnalis coming nearest to it in these particulars. e. NERVoUS SYSTEM. On account of the transparency of the body the central nervous system can be studied with ease in this species, either in the living animal or in whole preparations. In the ventral ganglionic chain there are, as in all species of Glossiphonia, twenty-one distinct ganglia. These innervate somites VII.—XXVII. respectively. The brain (cb., Figures 23, 27; also Figures 25, 26) represents the fused ganglia of the first six somites. The arrangement of its ganglionic capsules is the same as in G. stagnalis and G. fusca (Figures 8, 10, 12, 18). The two ventral capsules of somite vi. (6, 6, Figure 25) are arranged tandem, those of somites I.-v., side by side. The supra-cesophageal commissure lies well back, about over the lateral capsules of somite v. (Figure 26). 4. Glossiphonia heteroclita Liyyzxvus (1761). Plate 5; Plate 8, Figs. 35, 36, 38. Hirudo heteroclita Linneus (1761); H. hyalina O. F. Miiller (1774); Clepsine hyalina Moquin-Tandon (’26). a. Haprrat, Form, Size, Conor. This small and transparent leech is found both in Europe and in North America. Compared with G. stagnalis and G. fusca, it has a proportionally shorter and broader body (Plate 5, Figures 19, 22 ;, Plate 8, Figure 38. Com- pare Plate 1, Figure 1; Plate 2, Figure 4) ; in its movements, it is less active. It is found in ponds and sluggish streams, such as G. stagnalis frequents. Length of largest individuals, when extended, 13 mm. ; at rest, 8-9.5 mm. Width, extended, 3 mm.; at rest, 4.25 mm. Color. — The body is in general very clear and transparent, like that of a jelly-fish, but shows great individual variation in the matter of pigmentation. First, it always has more or less of a golden-yellow tint caused by the pres- CASTLE: NORTH AMERICAN RHYNCHOBDELLID&. 43 ence, in the deeper parts of the body, of large, rounded cells each containing a single yellow oil-drop, which is blackened when treated with osmic acid. Secondly, there are usually present (but this is the variable element in the pigmentation) irregularly rounded, oval, or even somewhat branched cells, which contain pigment granules either orange, dark-brown, or black in color. These cells are found near the dorsal surface of the animal, and often produce a conspicuous color pattern by their abundance in certain regions (Figure 38, Plate 8). In their finer structure, cells of this variety are rather closely related to the deep-seated pigment cells (reserve-food cells) found in G. stagnalis and G. fusca; but in respect to position (close to the surface), and occasionally in form (irregular or branched), they approach more nearly the superficial pig- ment cells (‘‘ excretophores,” Graf) of the species named. The pigmented areas which are often produced in G. heteroclita by the super- ficial pigment cells just described are (Figure 38, Plate 8), first, a median dorsal, longitudinal band, which, when best developed, extends, with occasional inter- ruptions, from about the seventh somite back to the anus. In the anterior ring of each somite it often broadens out into a trapezoidal form. Secondly, in about the same regions of the body (seventh to twenty-seventh somites), the anterior ring of each somite may be marked by a transverse, pigmented line, most conspicuous a short distance from the margin of the body, from which point it extends inward toward the trapezoidal, broad part of the median vitta, but rarely joins it. Apathy (88) has recognized as a distinct variety (striata) animals which have the transverse markings just described. It must be said, however, that one can find in a lot of animals collected from the same locality all gradations between forms with no pigment at all (of the superficial sort) and those having a median vitta and well-defined transverse striations. b. Rixes, Somites, Eyes, SucKERS, ETC. The surface of the body is rather smooth, being only slightly rougher than that of G. stagnalis. External rings, rather inconspicuous, particularly in the head region, where it is often difficult to determine their number and limits accurately. Number of preanal rings, seventy, counting as a single ring each of the so- mites I., II., XXVI., and xxvul., Figure 19. This number may be increased, if one counts subdivisions occasionally visible in some of the rings at the ends of the body. Somites I. and I1., as just indicated, are commonly uniannulate (Figures 35, 36, Plate 8); but somite 11. is sometimes subdivided by a transverse furrow (as shown in Figure 20, Plate 5). 1 Fat cells are also found in the deep parts of the body of G. stagnalis, G. fusca, and G. elongata, but the.contained oil-drops are in those species perfectly clear and transparent, so that they do not have the effect of pigment cells, as do the fat cells of G. heteroclita. : 44 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Somite 11., within the anterior part of which lies the mouth (or., Figure 20), is ordinarily biannulate, as are also somites lv. and xxv. (Figures 19, 35, 36). But in the section shown in Figure 20, ring 3, the anterior annulus of somite III., appears conspicuously subdivided, a rather unusual condition. On account of the obliquity of the section, the first three somites appear in that figure a little too long in proportion to their vertical dimensions. The sensilla shown in the anterior portion of ring 3 in Figure 20 is probably not one of the seg- mental sense-organs, for it is found on the wrong half of ring 3. Somites V.—xXIV. are triannulate, as in G. fusca, Somites xxvr. and xxvit. (reckoned as uniannulate) usually appear divided at the margin only into a broader anterior and a narrower posterior part. Compared with the species already described, the somite composition of G. heteroclita is about the same as that of G. fusca, somite abbreviation being less extensive in these species than in stagnalis and elongata. Eyes, usually six, the anterior pair small and generally, though not always, close together in ring 5 (Figures 35, 36, Plate 8). Sometimes this pair of eyes lies in ring 6; occasionally the pigment of one or both eyes is wanting altogether. The second and third pairs of eyes are most often found in rings 7 and 8 respectively, but one pair or the other or both may lie a little anterior or a little posterior to the ordinary position (compare Figures 35 and 36). The first and second pairs of eyes are directed forward and toward the side ; the third pair is directed backward and toward the side (Figures 20, Plate 5 ; Figures 35, 36, Plate 8). The eyes in this species seem to belong to somites Il., Iv., and v., respectively (Figure 20); but it is possible (though I think hardly probable) that a more careful study of the nerve connections would show that in this species, as in G. elegans (Figure 29, Plate 7), they have been derived from the sensille of somites 11.-1v. If so, the eyes have undergone a farther displacement backward in this species than in the case of G. elegans (compare Figures 20 and 29). Oral sucker, formed by somites I.-Iv. (Figure 20). Mouth (or., Figure 20), in the anterior part of somite m1., usually a little an- terior to the first pair of eyes. Posterior sucker, as in other species, slightly longer than broad (Figure 19). c. REPRODUCTIVE ORGANS. Male and female genital ducts open between the first and second rings of somite XII. (rings 28 and 29, Figure 19) by a common pore, a condition pecu- liar, I believe, to this species. Blanchard (’94) is certainly in error in describing the position of the genital pores as follows: “ Porus genitalis masculus inter annulos 25-26, vulva inter annulos 27-28 hians.” Testes (te., Figure 19), six pairs placed intersegmentally in somites XIII. XVIII. san The terminal part of the vas deferens (ejaculatory part) is un- XIV, XIX, CASTLE : NORTH AMERICAN RHYNCHOBDELLIDA, 45 usually stout and thick in this species and runs forward to the middle ring of somite XI. before turning sharply backward toward the genital pore (compare Figure 19 with Figures 4, 13, 27, and 28). The eggs, which in the vicinity of Cambridge are laid in May or June (at about the time G. fusca is laying), are whitish in color and are attached singly, not in groups as in the other species described, to the under side of the body (Figure 22). The eggs are of about the same size as those of G. stagnalis. The number laid varies greatly with the size of the individual, the observed extremes being eleven and sixty-five. Figure 22 shows in ventral view a large individual bearing forty-five eggs, each enclosed in a separate delicate sac which serves to attach it to the under side of the body. d. DiIGcEstivE TRACT. The mouth has the position most common in the genus, in the anterior part of somite 111. (Figure 20). The proboscis (pr’b., Figure 19) is long and the esophagus correspondingly short. The former ordinarily extends over somites Ix.—xu. and part of xII., and the latter ends in the anterior part of somite xiv., where the crop com- mences. The salivary glands (gl. sal., Figure 19) are large and distributed often through as many as seven or eight somites, usually somites XI.—XxvI1. The crop (?glv., Figure 19) bears six pairs of strongly developed lateral diverticula, a pair arising in the middle of each of the somites XIV.—XxIx. Some or all of the first five pairs may be bilobed distally, and each of the > sixth pair, which are very long, and extend back into somite xxu1I., bears about five secondary, lateral diverticula, which come off metamerically in somites XIX.—XXIII. The stomach (ga., Figure 19), with its four pairs of lateral diverticula, lies within somites XIX.—XXII. The intestine (in., Figure 19) begins about in somite xx. and extends back to the anus just behind somite xxviI. Proximally it consists of one or two chambers limited by valve-like constrictions. Posterior to this it gradually narrows backward. e. Nervous SYSTEM. The brain (cb., Figure 19) lies about in the eighth somite, The arrange- ment of its ganglionic capsules is peculiar in one respect. The ventral capsules’ of the last brain neuromere (Figure 21) lie side by side, not tandem as in the other species described in this paper. In other respects the arrangement of capsules is the same as that found in G. stagnalis and G. fusca (Figures 8, 12, 16,18). In the individual whose brain is represented in Figure 21, the most ventral and posterior capsule of neuromere I. had a horn-like process ex- tending back laterally into contact with the lateral cupsules of neuromere III. ; this condition, however, appears to be unusual. 46 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 5. Glossiphonia elegans Verriti (1872). Plate 7; Plate 2, Fig.53; Plate 3, Fig. 11. Clepsine elegans Verrill ("72); (?) C. pallida Verrill (’72); C. patelliformis Nicholson (’73). a. Hapirat, Size, Conor. This species is very closely related to the European G. complanata L. and G. concolor Apathy. Blanchard (’94), indeed, considers it identical with G. com- planata L. and regards G. concolor Apathy as merely a variety of the same species. However, both Apathy (’88) and Oka (94) testify to the perfect dis- tinctness of G. complanata and G. concolor, which occur together in Europe. I have myself compared animals of the species to be described with alcoholic specimens of G. complanata from Ziirich, Switzerland, and find certain small but constant differences between the two. I shall therefore describe the animals which I find here in the vicinity of Cambridge under the name proposed by Verrill in 1872, recognizing, however, that they are very closely related to the two European species (or varieties) named. G. elegans (Plate 7) is found in localities similar to those frequented by G. stagnalis, often in company with that species. It is considerably larger, being much broader and thicker in proportion to its length, though scarcely longer. In its movements it is more sluggish, resembling closely the small G. het- eroclita in that regard. It adheres to the side of the aquarium with a tenacity displayed by no other of our species except G. parasitica. The form of the body at rest is elliptical. The largest individuals which I have collected measure, when alive, as follows : — Length, fully extended, 28 mm. ; at rest, 14-18 mm. Width, fully extended, 5 mm. ; at rest, about 7 mm. Color. — Small individuals are usually of a bright, transparent green color. Adult animals, viewed with the naked eye or through a hand lens, appear of a reddish or greenish brown color, and are darker above than below. The head is colorless. The dorsal surface of the body is marked with numerous small circular white spots, about the width of a body-ring in diameter. These spots are so placed as to form transverse and longitudinal rows, just as do the similar spots of G. fusca, The transverse rows fall on the sensory (middle) rings of their respective somites, each row containing seven spots, when the full number is present. Each of these seven spots falls in a different longitudinal row, there being three pairs of rows arranged sym- metrically with reference to an unpaired (median) row, exactly as in G. fusca. The paired rows may be designated respectively paramedian, inter- mediate, and marginal, for they occupy practically the same position on the body as do the rows of white spots in the case of G. fusca, and the rows of papilla in that of G. parasitica (Figure 6). | CASTLE: NORTH AMERICAN RHYNCHOBDELLID. 47 In addition to the spots which fall into rows as just described, a few spots are usually found scattered more or less irregularly over the surface of the body. Two interrupted brown lines (Figure 30) appear in a paramedian position on the dorsal surface, the interruptions being due to the segmentally arranged white spots of the paramedian rows. A pair of similar, though fainter, dark lines is found on the ventral surface; but they are farther apart, including between them about the middle third of the ventral surface. The dorsal para- median lines include between them (in the middle of the body) about one fourth of the width of the dorsal surface, which part is usually rather more heavily pigmented than the more lateral portions. A median, clear, unpigmented band extends the entire length of the body on the ventral surface. The median row of light spots on the dorsal surface often run together in the posterior third of the body, forming a continuous light vitta. Examining more minutely into the coloration of the animal, one finds that it is due to the same two classes of cells as produce the coloration of most other species: first, pigment cells proper, — “ excretophores,” Graf; and secondly, reserve-food cells. The pigment cells proper, as in other species, occupy a superficial position in, or immediately underneath, the epidermis. They are stellate or richly branched, and are more abundant on the dorsal than on the ventral surface ; in small individuals they are almost entirely wanting. The pigment in immature animals is a rust-colored or dull reddish-brown, but in full-sized animals it is usually dark-brown. There is no pigment anterior and lateral to the eyes, nor in the white spots already mentioned. The pigment is more abundant than elsewhere in the paramedian dark lines, indeed its abundance there produces those lines. The reserve-food cells in this species, as in G. fusca, are of two forms: first, the ordinary form of large reserve-food cell distributed irregularly through the deeper parts of the body ; secondly, a special form of reserve-food cell, smaller, and more superficial in position, and found only in the white spots already described. The ordinary reserve-food cells are large and rounded in outline, often at- taining a diameter of forty mikra or more. They contain rounded granules of a bright green color both by reflected and by transmitted light. It is this form of cell which gives to the small, immature individuals their green color, and often imparts a greenish tone to the brown-colored adults. The special form of reserve-food cell agrees closely both in appearance and -in distribution with the similarly designated structures of G. fusca. It is found, as already stated, only in the white spots of the dorsal surface ; cells of this kind occur in a group of from two to a dozen or more each, situated in the centre of a white spot, just underneath the epidermis. By reflected light they are of a light lemon-yellow color ; by transmitted light, greenish-brown. VOL. XXXVI.— No. 2. 3 48 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Each of the white spots in the paired rows contains an inconspicuous, low rounded papilla (much less prominent than are the papille of G. complanata, so far as my observations go). The median row of white spots is less well developed than are the paired rows ; in the four or five somites immediately anterior to the anus, it is com- monly replaced by a continuous, median, clear vitta, within which is seen a narrower band of the lemon-yellow reserve-food cells. Obviously the color pattern of this species resembles very closely that of G. fusca, although in a majority of characters the animal is more closely related to G. parasitica. 6b. Surracr, Rines, Somites, Eyres, Suckers. The surface of the body is rather rough, owing to the strong development in this species of the integumental sense-organs described by Bayer (’98). It does not, however, bear conspicuous papille, as is the case with G. parasitica and the European G. complanata. The low, rounded papille which are found in the paired longitudinal rows of white spots are much smaller than the similarly placed papillae of G. complanata. In this particular G. elegans seems to agree with G. concolor (see Apathy, ’88, page 771). External rings, as a rule, rounded and distinct, less conver and not pointed as are those of G. complanata, sixty-eight in number, distributed as follows :— Somites I.—-Iv. uniannulate ; but the boundary between rings 1 and 2 is often inconspicuous (compare Figures 28, 29, 30), approaching the condition found in G. stagnalis, where somites I. and 11. form a single broad ring, which, how- ever, is sometimes divided by a shallow transverse furrow (Figures 3, 7). Somites v.-xxIv. triannulate, but the condition of somite v. is peculiar. Its anterior annulus (5, Plate 7, Figures 28-31) is commonly narrow and imper- fectly separated from the following (sensory) annulus (6). This case illus- trates well the initial step in reduction (or final step in elaboration, p. 33) of the triannulate somite. It represents an intermediate stage between the biannulate and triannulate condition of somite v. seen respectively in G. stag- nalis (Figure 7, Plate 3) and G. heteroclita (Figure 20, Plate 5). Somite xxv. is biannulate (Figure 28), but the furrow between its two annuli is often inconspicuous. Somites xXvi. and XxvII. are commonly uniannulate, though notched at the margin of the body, which fact shows that the final step in somite reduction (or initial step in somite growth) is not yet accomplished in the case of these somites. Eyes, six, in two parallel rows close together, in rings 3 and 4 (Figure 30). Sometimes the first pair of eyes lies partly in the posterior half of ring 2 (Figure 29). The middle pair is the latgest of the three ; the anterior pair, the smallest. The first two pairs are directed obliquely forward, the last pair obliquely backward; all are turned away from the median plane (Figures 29, 30). From the relation of the eyes to the nerves connected with the metamerie sensille (Figure 29), it is plain that the three pairs of eyes have been derived from the sensille of somites 1., 111., and Iv. respectively. It is further evi- Lae o> aed aE OE = CASTLE: NORTH AMERICAN RHYNCHOBDELLID. 49 dent that the single pair of eyes found in each of the species stagnalis, fusca, and elongata corresponds with the middle (largest) pair of eyes of this species, the pair belonging to somite III. The oral sucker, as in the other species described, lies within somites I.-Iv. (Figures 29, 31). c. REPRODUCTIVE ORGANS. Male genital pore (po. @, Figure 28), between somites x1. and x11. (rings 25 and 26), a position one ring anterior to that of the same structure in the species already described. Female genital pore ( po. 9, Figure 28), between the second and third rings of somite XI. (rings 27 and 28), the usual position of this structure in the genus. Testes (te., Figure 28), ten pairs. The anterior six pairs occupy the same positions as the testes in the species already described, being placed interseg- mentally in somites 7% _*YUt LY. XLK. behind those already mentioned ; the most anterior one, between the last . The remaining four pairs occur immediately crop and first stomach diverticulum, in somites pi ; the other three between RX: successive stomach diverticula, and like them separated by rather less than metameric intervals. No other species of Glossiphonia known to me, except the European G. complanata, has normally a greater number of testes than six pairs. In that species likewise the testes number ten pairs placed exactly as in elegans. This is one of several facts showing the very close relationship of the two species named. The last one or two pairs of testes are less constant in their occurrence than those farther forward. Eggs are laid by G. elegans, in the vicinity of Cambridge, in April, May, or as late as June. The temperature of the water in the spring undoubtedly exercises considerable influence in determining the time of egg-laying. Indi- viduals brought into the laboratory on March 27, 1898, laid eggs nine days later. On April 29, 1900, animals of this species bearing eggs were collected from Alewife Brook, Cambridge, though G. stagnalis, found with them, appar- ently had not yet laid its eggs. The eggs are dull pinkish white in color and are borne on the under side of the body in from three to six large clusters, which are rather easily detached from the body, if the animal is disturbed. Each cluster contains a considerable number of eggs, often as many as twenty or twenty-five, enclosed in a delicate sac. The sacs are not arranged sym- metrically in two parallel rows, as in G. stagnalis and G. fusca, but quite irregularly, a sac being attached either in the median plane of the body or to one side of it, as the case may be. d. DicEstivE TRActT. The mouth is situated well forward in somite 111., anterior to the eyes, or at least anterior to the last two pairs of eyes (Figures 29, 31). 50 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. The proboscis (pr’b., Figure 28) is long, extending over somites VIII.—XII. There is practically no esophagus, as I have used the term, for the pharyngeal sac containing the proboscis extends back almost to the beginning of the crop. The salivary glands are numerous, often reaching seventy-five or more in number in each half of the body. They are scattered usually through somites XI.-xvill. In Figure 28 they are represented as relatively a little too small. The crop (?glv.) bears seven pairs of large, lateral diverticula directed back- ward and often lobed distally. They arise in somites x11I.—XIX., always in the middle of a somite, as in the other species described. The last pair of crop diverticula is, as usual, the largest of all; it may extend back through three or four somites, giving off secondary lateral diverticula metamerically, as shown in Figure 28. Often, however, when the crop is empty, the last pair of diver- ticula is little longer than the preceding pair. The stomach (ga., Figure 28) bears, as in other species, four pairs of diverti- cula, which arise within the three somites x1x.-xx1. The intestine (in.) extends through the six remaining somites, consisting proximally of two distinct cham- bers limited by valve-like constrictions and usually situated in somites XXII. and xx. Distally it is a gradually narrowing tube terminating at the anus just behind somite xxvil. e. NEPHROPORES, NERVOUS SYSTEM. The nephropores open ventro-laterally, a little anterior to the middle of the sensory ring of a somite. The number of nephridia has not been determined for this species. The brain (cb., Figures 28, 30) lies for the most part in somite vil. The arrangement of its ganglionic capsules (Figure 5, Plate 2; Figure 11, Plate 3) is usually similar to that found in the brain of G. stagnalis and G. fusca, but the capsules are not so closely crowded together, and the supra-cesophageal com- missure lies well forward, not being carried back over the middle of the brain as in G, stagnalis (Figure 12). The less crowded condition of the capsules in this species (Figure 5) explains an abnormality in their arrangement observed in the brain of a single individual out of several examined; the two ventral capsules of somite 111. (usually found side by side as in G. stagnalis and the other species already described) were in this case arranged tandem, just as in ganglia in unabbreviated somites. Comparing the conditions of the brain capsules in the several species described in this paper, one may say that the larger the leech is, the less are its capsules crowded. This fact seems to indicate that the capsules, and probably the indi- vidual ganglion cells also, do not increase in size proportionally with the growth of the leech. This is certainly true of the development of the individual, if not also of the race, for in the very young leech the ganglia of the nerve chain oc- cur in close succession with scarcely any intervening space, whereas in the adult they may be separated by a distance of two rings or even more, CASTLE: NORTH AMERICAN RHYNCHOBDELLIDA. 51 6. Glossiphonia parasitica Say (1824). Plate 1, Figs. 2, 3a, 36; Plate 2, Fig. 6; Plate 8, Figs. 32, 33, 37. Hirudo parasitica Say ('24) ; Clepsine parasitica Diesing (50); C. plana Whitman (912); ¢ C. chelydra Whitman (’914). a. Hapitat, Form, Size. This large and conspicuously colored leech is the commonest and most widely Bistributed of our North American species of Glossiphonia. It is often found adhering to the bodies of turtles, whose blood it sucks, or underneath stones in pools and streams frequented by turtles. It is referable to the genus Placobdella Blanchard (’94), if one recognizes the validity of that genus. In it are included probably several forms which because of their close relationship I choose to call varieties. One of these has been carefully described by Whitman (91%) under the name “Clepsine plana.” In what follows I hope to supplement that de- scription and add the description of another form which is commonly found associated with it. The two varieties agree completely, so far as I can deter- mine, in form, size, and constitution of somites, but can be distinguished in my collections by constant differences in roughness of surface and in color pattern. In general form the body in this species is very broad and flat. Whitman - describes it correctly in the case of large individuals as “ ovate-elliptical in con- traction, emarginate posteriorly.” In the case of small individuals, however, or of large individuals well extended, the emarginate condition is not present (Figure 6, Plate 2; Figure 37, Plate 8; Figure C, p. 56). The dimensions given by Whitman for the largest individuals, I can substantiate: “ Length at rest, 5-6 cm.; width, 2.6 cm.” I have an alcoholic specimen (var. rugosa) from Lake Chautauqua, N.Y., which measures 5.6 cm. in length, and 3 cm. in width. Another (var. plana) taken from a turtle brought from the Illinois River measures 5.5 cm. in length, 2.3 em. in width. A living specimen (var. plana) taken from a snapping turtle (Chelydra serpentina) captured near Cambridge, Mass., measures at rest 5.8 cm. in length, 2.1 cm. in width. Whitman says further ; ** Length in extension, 8.5cm.; width, 1.8 cm.” My living Cambridge specimen attains in extension a length of about 7.5 cm., in which condition its greatest width is 1.5 to 1.7 cm. 6. Riyxes anp Somirss. The rings are distinct except at either end of the body. The furrow between the anterior and middle rings of each somite is, however, less deep than that which separates other rings, for which reason the anterior two thirds of a so- mite sometimes appears like a single broad annulus, especially at the margin of the body (Figures 2, 36, Plate1; Figure 6, Plate 2; Figures 32, 33, 37, Plate 8). Somites 1., ., and XX V.-XXVII. uniannulate (Figures 6, 33, 37), but xxv. and 52 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. XXVI. are commonly divided at the margin of the body into a broad anterior and a narrow posterior portion. Somites 111. and Iv. are biannulate, the broad anterior ring in each case bearing the sensillz and representing both the an- terior and the middle ring of a triannulate somite (Figure 2, IlI.-vI.). The remaining preanal somites (V.—xxIv., Figure 6) are triannulate, but the pos- terior annulus of XxIv. is narrower than the adjacent annuli (Figure 6), and the anterior and middle annuli of somite v. are united ventrally while sepa- rated by only a very shallow furrow dorsally (7, 8, Figures 2, 36, Plate 1. These two cases illustrate the centripetal progress of abbreviation (or arrested development), that part of each terminal triannulate somite being affected which is adjacent to an abbreviated somite. In Figure 32, Plate 8, is shown a rather unusual condition, the apparent disappearance of the furrow separating somites If. and 11.1 The total number of preanal rings is sixty-nine, counting somites I., IL, and XXV.—XXVII. as uniannulate, Il. and Iv. as biannulate, and v.-xxIV. as triannulate (Figure 6). ce. Eyres, Mourn, Ora Sucker. The eyes appear in the living animal, or in whole preparations, as a single pair closely united and situated in rings 3 and 4 (somite 11.). See Figure 6, Plate 1; and Figures 32, 33, Plate 8. An examination of sections, however, particularly of young individuals, shows that there are really three distinct pairs of eyes present, there being a small rudimentary pair anterior, and an- other still more rudimentary posterior to the principal pair of eyes, exactly as shown for “C. hollensis” by Whitman (’92, Figure 6). All three pairs of eyes? are partially imbedded in a common pigment mass, the anterior and middle pairs being directed forward, the posterior pair back- ward, just as in G. elegans and G. heteroclita (Figures 20, 29). The largest 1 A similar condition is figured by Whitman (’91®) in his Plate 15, Figure 1. In his text, however, Whitman says (p. 412): “In front of the eyes I was unable to discover any distinct rings. In another species C. chelydre, from Wisconsin, there are three narrow rings in front of the eyes; and the first is marked by the usual metameric sense-organs. Although no metameric sense-organs were recognized in front of the eyes in C. plana, the correspondence of other metameric characters in the two species is sufficiently close to enable me to identify the ocular rings as equivalents. The preocular part of the head is, therefore, probably equivalent to the first somite of C. chelydre, and is so numbered in Figure 1.” In view of Whitman’s subsequently published studies on “ The metamerism of Clepsine” (92), I think he unquestionably would now recognize two preocular somites both in “C. plana” and in “C, chelydra ”; at any rate, that is the number found in the species which I am describing (Figure 2, Plate 1), Since Whitman has pointed out no other difference between his “plana” and “ chelydre ” than the uncertain one of preocular rings, I consider that their specific distinctness remains to be established. ? Only the largest (middle) pair of eyes appear in the section shown in Figure 2. . _— — yo CASTLE: NORTH AMERICAN RHYNCHOBDELLIDAZ. 53 (middle) pair is closely united with sensille situated in the first ring of so- mite 111. (Figure 2), a fact which Whitman (’92) established for ‘‘ C. hollensis” and which I can completely confirm for the species under discussion (Figure 2),} Whitman (92) further established the fact that the anterior pair of eyes in “ hollensis”’ originates in connection with the sensille of somite mu. He gives no statement as to the origin of the posterior pair. Comparison with G. elegans (Figure 29), however, leads me to regard this pair as probably derived _ from the sensillz of somite Iv. If so, the condition of the eyes in parasitica can be derived in its entirety from that found in G. elegans by supposing that both the anterior and the posterior pairs of eyes have become rudimentary and been brought close to the large middle pair. The mouth (or., Figure 2) apparently lies between somites I. and 1. ; in other species it lies farther back, usually in the anterior part of somite 11. The oral sucker is formed by somites I.-Iv., as in other species. d. REPRODUCTIVE ORGANS. The genitaz pores are situated in this species exactly as in G. elegans; the male (po. @, Figure 36), between somites xI. and x. (rings 27 and 28) ; the female (po. 2), between the middle and posterior annuli of somite x11. (rings 29 and 30). XII. XVIII. Testes, six pairs situated intersegmentally in somites ——— _ , the usual REVS) (Ex position in the genus. The eggs are large, white, and opaque. In the vicinity of Cambridge they are laid in May and June, perhaps also in July. In the case of those animals which laid in the laboratory, the eggs appeared to be attached loosely in a sin- gle group of fifty or more to the side of the aquarium, rather than to the body of the leech as is the case in the other species studied. The leech remained closely arched over the eggs, —a position from which it was removed only with great difficulty. e. Digestive Tract. The digestive tract resembles very closely that of G. elegans, but has one strikingly distinctive feature : the salivary glands (gl. sal., Figure 30), instead of being distributed through several somites in the crop region, are closely aggregated into two compact groups in each half of the body, these groups lying symmetrically, a pair on either side of the proboscis, within somites rx.-xI. 1 On account of this and other close structural agreements with “ C. hollensis ” as described by Whitman (’92), I was for some time inclined to regard that name as well as “chelydre ” as a synonym with parasitica, and I have so treated it in a recent publication (Castle, 1900). Professor Whitman, however, has subsequently informed me in a letter that in hollensis “there are several pairs of pigmented eyes behind the pair usually recognized as ‘eyes.’ These are quite conspicuous in the living leech, and I have never seen any such feature in other Clepsines.” This being so, it is probable that hollensis should rank as a distinct species. 54 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. The crop bears seven pairs of lateral diverticula, as in G. elegans and the closely related European G. complanata, with both of which this species has many points in common. ‘The first pair of diverticula arise in the anterior or middle part of somite x11. and are two or three lobed, the anterior lobe being prolonged forward through somites xu. and x1. The five following pairs of crop diverticula arise in the middle of somites x1v.-xvmt. respectively, and are usually bilobed distally. The last (seventh) pair of crop diverticula ex- tend far back of their origin in somite x1x., often into somite xxm. They give off secondary lateral diverticula, a pair in each of the somites through which they extend. The crop diverticula are often a conspicuous feature of this species when viewed in a living condition from the ventral side of the animal, for numerous large green chromatophores aggregate about the crop and show through the clear ventral body wall the form of the crop outlined in green. f. Neparopores, Nervous System. The nephropores (nph’po., Figure 36) open ventrally, anterior to the middle of the sensory ring of a somite, as stated by Whitman (91*). They are present in the eighth and all the following triannulate somites. I have nothing new to add to Whitman’s (’92) excellent account of the cen- tral nervous system. ‘It is important to notice, however, the arrangement of the ventral capsules in the brain region (Figure 36). Those of neuromeres 111.—vI. all lie in a single row in the median plane; that is, have what I have called the tandem arrangement. The ventral capsules of neuromere 11. (2, 2, Figure 3 0) have the side-by-side position found in all the species examined by me. Figure 3 a is a dorsal view of the brain and shows that the supra-csophageal commissure in the species lies far forward in what may well be regarded as its primitive position. The less crowded condition of the brain capsules in this as compared with other species is interesting, as showing that the smaller the leech is, the more crowded are its brain capsules likely to be (compare page 50). g. PAPILL&, COLORATION. I have reserved to the last, in describing this species, the discussion of papil- le and coloration, for it is on the basis of these characters alone that I am able to distinguish two varieties, plana and rugosa, which I find associated together, but apparently without intergrading forms, in collections from Cambridge, Mass., Lake Chautauqua, N. Y., Lake Forest, Ill., and Wellsville, Kan., avery wide range extending across the Mississippi valley and the Atlantic seaboard. (1) Var. plana Clepsine plana Whitman, ’91*). This variety has a relatively smooth skin, which bears dorsally small dome- shaped papillae, the most conspicuous of which are placed as indicated by stars ‘ CASTLE: NORTH AMERICAN RHYNCHOBDELLID#. ae in Figure 6, Plate 2. They include five longitudinal rows of papille found on the middle (sensory) annuli of usually all the triannulate somites. These rows may be designated, from their position, median, marginal, and intermediate, the first named being unpaired, the other two paired. A row of papille is found also between the median and each intermediate row, but these papille are situated not on the middle, but on the posterior annulus of each of the somites from about vim. to xxtl. inclusive (Figure 6). These will be designated paramedian rows. The most conspicuous papille of somites xxv.-xxvu. are usually placed as indicated in Figure 6. They consist, first, of a continuation of the marginal rows back to the anus; secondly, of two rows of three papillz each, placed one on either side of the median plane and too near it to fall in the paramedian rows found farther forward. Other less conspicuous papille occur on the dorsal surface of the body and posterior sucker, but no papille are found on the ventral surface of the animal. The general color of the body above is brown variegated with yellow, orange, and green. Light areas of yellow or pale orange form : — I. A median vitta extending from the anterior end of the body back to somite xxy., usually without interruption, but not always so, and expanding commonly at six places, namely, (1) in somites vi. and vi. (Figure 32, Plate 8) ; (2) in somite 1x.; (3) in somites xu. and xi.; (4) in somites xv. and xvi. ; and (6) in somite xxi. (and the posterior part of somite xx1.). The median row of papillz already described falls entirely in the median light vitta. In somites Xxv.-xxvul. both vitta and papille become double, dark pigment being found along the median line back to the anus, usually behind it also quite to the posterior margin of the acetabulum. The double (or paramedian) light vitta of somites XXV.—XXVII. contains the three pairs of papillz shown in Figure 6, Plate 2 ; it may or may not be continuous with the median light vitta farther forward. Il. Throughout the greater part of the body the papille of the intermediate rows lie each in an irregularly rounded light spot. The successive spots of each half of the body may become confluent so as to form an irregular, frequently interrupted, longitudinal band. Ill. The margins of the body are conspicuously marked with metameric light spots from about the third or fourth somite back to somite xxv. Some idea of the form and position of these spots may be obtained from an examina- tion of the stippled areas in Figure 6, Plate 2, and Figure 32, Plate 8. Each spot is typically V- or U-shaped and is placed on the adjacent non-sensory rings of two successive somites. The usually hollow centre of the V or U is formed by a spot of brown sometimes bordered with orange. The margin of the sen- sory ring is generally darker in color than its more median parts, so that it is strongly in contrast with the metameric light spots which it separates. The posterior sucker is decorated with radially placed triangular light spots (Figure 6) resembling the marginal spots of the body. Other irregularly 56 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. rounded light spots may be found on the dorsal surface of the body in light- colored individuals, usually associated with certain papille. There is a certain correlation in the development of light spots in different parts of the body ; an animal which has a well-developed median vitta will also have conspicuous marginal and intermediate light spots and vice versa. The ventral side of the body is much lighter in color, marked only by a few longitudinal bands of dull brown or greenish brown. The number of these bands is either eleven or twelve according as there is present, in addition to five pairs of bands laterally placed, a single broad median band or a pair of narrow paramedian bands separated by an irregular median clear band. From the under side of the body one can often see in living animals the green pigmented crop diverticula showing through the semi-transparent body. XXV. XXVI. XXVII. Ficure C.— G. parasitica, var. rugosa. Dorsal view of posterior part of body, showing position and approximate relative size of papille. From a Cam- bridge, Mass., individual. (2) Var. rugosa, var. nov. The dorsal surface of the body is much rougher in this variety, the papille being larger, more numerous, and structurally more complex. Instead of being simple, low, and dome-shaped, the more conspicuous papille are extended dis- tally in several divergent whitish points, giving the body a decidedly rough, harsh feeling to the touch in the case of hardened specimens. The larger papille are likewise rendered more conspicuous by the fact that they are commonly unpigmented, though placed in a generally dark background. The arrangement of the principal rows of papille on the dorsal surface is similar to that in G. plana, but with the following easily determined and con- stant difference. In somites xxi. and xxtv. (Figure C), the median row of papilla becomes inconspicuous or disappears altogether, and a large papilla appears on either side of the median line, on the sensory ring of each somite. The ventral surface is free from papilla as in plana. CASTLE: NORTH AMERICAN RHYNCHOBDELLID#. 57 The color pattern is somewhat similar to that of plana, but the contrasts are less striking and the colors less brilliant. The general color effect of the dorsal surface is a grayish brown. Marginal spots of light yellow are present, as in plana, on the non-sensory rings, but they are smaller and do not extend so far mesiad from the margin of the body. Practically all the larger papille appear as small white spots in a generally dark background. The median vitta is not a continuous light band as in plana, but is inter- rupted at regular intervals by spots of a darker color than the general dorsal surface. It begins as a narrow median light band on the head and neck, con- stricted or sometimes interrupted in the posterior part of somite v1., less often constricted or interrupted in somite v. also. About in annulus 19, somite Ix., begins a narrow dark band which continues to the middle of somite xtt. Then come alternating light and dark spots, three of each. A light spot ex- tends over four annuli, a black spot over five as follows: Light spots, annuli 29-32 (Figure 6), 38-41, 47-50; dark spots, annuli 33-37, 42-46, 51-55. An- other light spot covers rings 56-64 or 65, broadening out posteriorly so as to include the paired papille of somites xxi. and xxiv. (Figure C). This is fol- lowed by a median dark spot extending back past the anus to the margin of the posterior sucker. The posterior sucker is marked by alternating light and dark rays, very much as in plana (Figure 6); it also bears papille like those of the body farther forward. Ventrally the body is light gray in color, owing to the presence there of scattered pigment flecks, which, however, are not arranged in longitudinal bands as in plana. V. MUTUAL RELATIONSHIPS OF THE SPECIES DESCRIBED. The species described in this paper, with the exception of heteroclita, fall naturally into two distinct groups (Figure D, page 58), which may be designated respectively the stagnalis and the parasitica groups. The former includes the three species stagnalis, elongata, and fusca; the latter, parasitica and elegans, with the closely related European species, complanata and concolor. Heteroclita occupies a somewhat isolated position intermediate between these two groups. As arranged in Figure D., the species form a series in which there is from left to right an increasing degree of complexity of structure. This appears from an examination of rugosity, somite structure, crop diverti- cula, and certain other characters. In the species of the stagnalis group (1) there is a single pair of eyes derived from the sensillze of somite 111., (2) the genital pores are separated by a single ring, namely, the middle (sensory) ring of somite xu, and 58 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. (3) the crop diverticula are simple and never exceed six pairs in number. (4) All three species are small, (5) have relatively smooth skin, and (6) at least two of them bear the eggs in clusters attached symmetrically in a double row to the under side of the body, the condition in the third species being unknown. In parasitica and elegans (1) there are three pairs of eyes derived respectively from the sensilla of somites U., 111, and Iv., (2) the genital pores are separated by two rings, the anterior two rings of somite XII., (3) the crop diverticula number seven pairs and are lobed, (4) the in- tegument is rough and bears papille, (5) the attachment of the egg heteroclita ie elegans stagnalis | concolor elongata complanata Zageye parasitica / Fictre D. Diagram indicating relationships of the species described. clusters to the body, when such attachment exists, is imperfect and the arrangement of the clusters irregular. | The European species complanata and concolor are very closely related to elegans, complanata certainly, perhaps also concolor, being intermediate between it and parasitica. In view of the many points of similarity between parasitica and com- planata, there seems to me to be insufficient ground for placing them in distinct genera, as proposed by Blanchard. Allusion has already been made to the somewhat isolated position of heteroclita. In size and in the character of its integument, it resembles the stagnalis group, likewise in the number of its crop diverticula ; in regard to the lobed condition of its crop diverticula, it resembles the parasitica group. In the number of its eyes (three pairs), it likewise resembles the latter group, but the derivation of these apparently is from different somites (111.-v. in heteroclita, 11.—1v. in parasitica and elegans.) As regards the position of the genital pores and the way the eggs are borne, it differs alike from both groups. CASTLE: NORTH AMERICAN RHYNCHOBDELLIDA. 59 VI. PARASITES. Three different endo-parasites, of which I find no notice in the litera- ture, in addition possibly to one already described by Bolsius (’96), infest more or less commonly the species of Glossiphonia found in the vicinity of Cambridge, Mass. One of these is a small nematode, another a trematode, these two having been observed in the body of G. stagnalis only ; the third is a sporozodn found in at least four of the species described in this paper. In January, 1898, I first observed a minute xematode parasite wrig- gling about in the central lacunar space of a live G. stagnalis. Another similarly parasitized leech was found upon further search, and a third was found in the following March, the ovary of the host containing at that time full-grown eggs. The parasite in the last-mentioned case lay close to the contractile dorsal blood-vessel, a very common position for it, as subsequent observations showed. In the spring of 1899 several parasitized individuals were collected and studied; and others were observed in the fall of 1899. The length of the parasite is about the same in the case of all individuals examined ; namely, 1.43 mm. In form, the worm is slender and thread-like, being widest near the middle of its body, where it measures 0.027 mm. in breadth. From there it tapers almost imper- ceptibly toward either end. The posterior end of the body is sharply pointed ; the anterior end blunt, its centre being occupied by the very minute, conical mouth. Examination of a large number of individuals of G. stagnalis in the spring of 1899 showed that between five and ten per cent of the indi- viduals taken from a particular pond, in which the species abounds, contained the nematode parasite. Usually only a single parasite has been observed in the body of a host, but in one case there were three. The nematode is generally found either coiled up (but not encysted) or wriggling about in the central lacuna (body cavity), in the middle or toward the posterior end of the body. The presence of the parasite does not seem seriously to inconvenience its host, for the parasitized. individuals are as large and well developed as those free from parasites, and contain sexual products in equal abundance. Parasitized individuals were kept in aquaria for several weeks without the occurrence of any noticeable change in the condition of the parasites. This fact and the manifest immaturity of all the parasites examined makes me believe that the leech is an intermediate host and that the 60 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. nematode probably attains maturity after passing from the body of the leech into that of another host, perhaps some fish, which feeds upon the leech. How the nematode gets into the body of the leech is likewise unknown, probably from the body of some snail or other small pond animal on which the leech feeds. The supposed trematode parasite I have observed but once, in Novem- ber, 1899, when three individuals were observed encysted in a single G. stagnalis. Unfortunately they with their host died in captivity before I had an opportunity to study them carefully. They lay imbedded in the deeper muscle layers of their host’s body, toward its anterior end, each enclosed in a delicate rounded cyst. A single ventral sucker was observed in the parasite and this seemed to lie a little nearer one end of the body. Toward the opposite end, a dark granular substance was observed in the interior of the body, probably in the digestive tube. My study of the parasite, was so incomplete that I should not feel war- ranted in asserting the absence of a second sucker more nearly terminal in position than the one observed. No measurements of the cysts were made, but I should estimate their diameter roughly at 0.50-0.75 mm, About half of the individuals of G. elongata which have come under my observation contain a gregarine which appears to be identical with that described by Bolsius (’96) as occurring in G. complanata (Clepsine sexoculata). I have not, however, made a sufficiently careful study of it to enable me to add anything to his account. I find the parasite attached always to the wall of the stomach diverticula (Figure 27, ga.), never in crop or intestine. A majority of the individuals of G. fusca collected by me contain sporozoa in an encysted condition. These parasites are quite common also in the body of G, heteroclita and that of G. elegans, and I have found them in a single individual of G. stagnalis. Whether or not they represent another stage of the gregarine found in G. elongata, I am unable to say. As already indicated, I have ob- served them only in stages of encystment, more or less advanced. One finds the heavily staining sporocyst in whole preparations of its host, usually near the margin of the body, imbedded in the deeper-lying muscle layers (longitudinal and dorso-ventral). The sporocysts which I have observed were spherical in form ; the largest ones examined were about 0.13 mm. in diameter and were protected by a thick, dense wall. I have not yet been able to obtain sporocysts containing fully formed CASTLE: NORTH AMERICAN RHYNCHOBDELLIDZ. 61 spores. ata, accordingly, are wanting for a full description of this parasite, as well as of the others mentioned, and only a portion of its life history is known. Nevertheless I insert this notice in the hope that some one else may be able hereafter to make use of my fragmentary observations. CamBrineGs, Mass., June, 1900. 62 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. BIBLTOGHRALPILY. Apathy, S. °88. Siisswasser-Hirudineen. Ein systematischer Essay. Zool. Jahrb., Abth. f. Syst., Bd. 3, pp. 725-794. Apathy, S. °88°. Analyse der ausseren Kérperform der Hirudineen. Mitth. Zool. Sta. Neapel, Bd. 8, pp. 153-282, Taf. 8, 9. Bayer, E. $8. Hypodermis und neue Hautsinnesorgane der Rhynchobdelliden. Zeit. f. wiss. Zool., Bd. 64, pp. 648-696, Taf. 23-25, 10 Textfig. Blanchard, R. °94. Hirudinées de I’Italie continentale et insulaire. Boll. Mus. Zool. ed. Anat. comp. Torino, Vol. 9, n. 192, 81 pp., 80 Fig. Bolsius, H. 96. Un parasite de la “Glossiphonia sexoculata.” Mem. Pontif. Accad. Nuovi Lincei, Vol. XI., 5 pp., 1 Pl. Bristol, C. L. °99 The Metamerism of Nephelis. Jour. of Morph., Vol. 15, pp. 17-72, Pls. 4-8. Budge, J. ‘49. Clepsine bioculata Savigny. Verh. d. naturh. Vereins preuss. Rhein- lande, Bd. 6, pp. 89-155, Taf. 5, 6- Castle, W. E. 1900. The Metamerism of the Hirudinea. Proc. Amer. Acad. Arts and Sci., Vol. 35, pp. 285-303, 8 Fig. Diesing, K. M. °50. Systema helminthum. 8vo, 2 vol., Vindobonae. Graf, A. °99. Hirudineenstudien. Abhandl. Leop.-Carol. Akad. d. Naturf., Bd. 72, Nr. 2, pp- 217-404, Taf. 1-15, 26 Textfig. Gratiolet, P. 62. Recherches sur lorganisation du systéme vasculaire dans la sangsuc médicinale et ’aulostome vorace. Ann. Sci. Nat., Zool., T. 17, pp. 175-225, ig BA Johnson, J. R. ‘16. Observations on the Hirudo vulgaris. Phil. Trans. Roy. Soc. London, pp- 13-21, Pl. 4. Lee, A. B. 96. The Microtomist’s Vade-mecum. Fourth edition. Philadelphia. CASTLE: NORTH AMERICAN RHYNCHOBDELLIDA. 63 Linnaeus, C. 1758. Systema Naturae. 10th edition. Moore, J. P. °98. The Leeches of the U. S. National Museum. Proc. U.S. Nat. Mus. Vol. 21, pp. 543-563, Pl. 40. Moore, J. P. 1900. A Description of Microbdella biannulata, with Especial Reference to the Constitution of the Leech Somite. Proc. Acad. Nat. Sci. Philadel- phia, pp- 50-75, Pl. 6. Moquin-Tandon, A. 46. Monographie de la famille des Hirudinées. 8vo, 448 pp. Atlas, 14 Pl. Paris. Miller, O. F. 1774. Vermium terrestrium et fluviatilium, etc. 4to. Vol. 1, Havniae et Lipsiae. (See Vol. I., “ Helminthica,” p. 49.) Nicholson, H.A. | 73. Contributions to a Fauna Canadensis; being an Account of Animals dredged in Lake Ontario in 1872. Canadian Journal, Vol. 13, pp. 490- 506, 4 Textfigs. Oka, A. °94. Beitrage zur Anatomie der Clepsine. Zeit. f. wiss. Zool., Bd. 58, pp. 79-151, Taf. 4-6. Savigny, J. C. 20. Systeme des Annélides. Folio, Paris. Say, T. °24. Zodlogy. Keating’s Narrative of an Expedition to the Source of St. Peter’s River, etc., in 1823, under 8S. H. Long. 2 vols. Philadelphia. Vol. 2, pp. 253-378, Pl. 14, 15. Verrill, A. E. °74. Synopsis of the North American Fresh-Water Leeches. Rept. U. 8. Fish Commissioner for 1872-3, Pt. 2, pp. 666-689. Whitman, C. O. : 85. The External Morphology of the Leech. Proc. Amer. Acad. Arts and Sci., Vol. 20, pp. 76-87, 1 Pl. Whitman, C. O. *91. Spermatophores as a Means of Hypodermic Impregnation. Jour. of Morph., Vol. 4, pp. 361-406, Pl. 14. Whitman, C. O. 91% Description of Clepsine plana. Jour. of Morph., Vol. 4, pp. 407-418, PRtl5- Whitman, C. O. *92. The Metamerism of Clepsine. Festschr. Leuckarts, pp. 385-395, Pl. 39, 40, Leipzig. 64 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. EXPLANATION OF PLATES. All figures were drawn with the aid of Abbé’s camera lucida, unless otherwise stated in the explanation of figures. Arabic numerals in the figures designate rings, which, except in the case of Fig- ures 23 and 27, Plate 6, are numbered from the extreme anterior end of the body backward; Roman numerals designate somites numbered in the same manner. ABBREVIATIONS. act. Acetabulum (posterior sucker). oc. Eye. an. Anus. a. C&sophagus. cb. Brain. or. Mouth. dt. e). Ejaculatory duct. po. & Male genital pore. ga. Stomach. po. 2 Female genital pore. gl. d. Dorsal gland. pr. Proboscis. gl. sal. Salivary glands. sac, phy. Pharyngeal sac. in. Intestine. suc. or. Oral sucker. gle. Crop. te. Testis. lac. marg. Marginal lacuna. va. df. Vas deferens. npl’po, | Nephropore. va. ef. Vas efferens. oa. Ovary. vs. sem. Seminal vesicle. CasTLE. — Rhynchobdellidz. PEAGE oa Fig. 1. G. stagnalis. Entire digestive tract shown; somite limits indicated by transverse lines, rings not represented. From an entire preparation. X about 16. Fig. 2. G. parasitica. Parasagittal section of head end of a small individual taken from a turtle (probably Chelopus insculptus Le Conte) bought in a Philadelphia market. Only one (the largest) of the three closely asso- ciated pairs of eyes appear in the section. Fig. 8. G. stagnalis. Ventral view of head end, showing mouth, oral sucker, and the marginal sensille and annulation of somites 1—v1. From an entire preparation. X 83. Fig. 3a. G. parasitica. Dorsal view of brain. Fig. 3). G. parasitica. Ventral view of anterior part of a small individual obtained from the same source as that shown in Figure 2. From an entire preparation. _#® PuaTe L ZA Senet irae e With, ace Cee ae ee a : 7 — : ee —————————— eee ae Uh ee eee ee eee CasTLE. — Rhynchobdellide, PLATE 2. Fig. 4. G. stagnalis. Diagram showing annulation, central nervous system, repro- ductive organs (male in left, female in right half of figure), nephropores, ete. The outline of the body was drawn from a whole preparation (X about 16); everything else is diagrammatic, representing the average form and position of organs as determined by examination and com- parison of several individuals. Fig. 5. Brain of G. elegans, ventral view. From an entire preparation. X 42. Fig. 6. G. parasitica. Dorsal view of a young individual from Havana, Illinois, partially extended. ™X about 10. The starlike structures indicate papille ; not all of those shown were observed in the individual figured, some being supplied from the study of larger individuals in which the papillz are more conspicuous. PLATE 2. = | Vicor CE peed I oom Ht ts B. Meisel, lith Boston + CASTLE. — Rhynchobdellide. PLATE 3. Fig. 7. G. stagnalis. Parasagittal section of anterior part of body. X 96. Fig. 8. G. stagnalis. Brain viewed from left side. Reconstructed from sections. > 208. Roman numerals designate segmental nerves; Arabic numerals, the ganglionic capsules which supply nerve fibres to same. Fig. 9. G. stagnalis. Posterior part of ventral ganglionic chain, dorsal view, reconstructed from frontal sections. Arabic numerals designate ven- tral ganglionic capsules; Roman numerals, metameric nerve bundles. x 170. Fig. 10. G. stagnalis. Diagram showing the arrangement of ganglionic capsules on the ventral surface of brain. Fig. 11. G. elegans. Dorsal view of brain. Fig. 12. G. stagnalis. Dorsal view of anterior part of brain. From frontal sec- tions combined. X 167. Piatt" 3: -RHYNGHOBDELLIDAE.. a saky B. Meisel, lith. Boston. ae =F Se.) te Fig. Fig. Fig Ss Fig. Fig. Fig. CasTLE. — Rhynchobdellide. 18. PLATE 4. G. fusca. Dorsal view of a small individual. For clearness furrows between annuli are represented only at the margin of the body, except where they mark somite boundaries. Testes are shown only in the right half of the figure, salivary glands only in the left half. From an entire preparation. X about 34. Parasagittal section of head end. X 82. Head end of young individual viewed from left side. From an entire preparation. X 83. Head end of individual shown in Figure 13. Dorsal view. X 83. Group of reserve-food cells from one of the segmental clear spots marking the sensory annuli. From a living animal. X 368. Ventral view of brain. From an entire preparation. X 208. & = a Pre “— = | 3 = a PLATE 5 E eo => ores ar CastLE. — Rhynchobdellide. Fig. 19. Fig. 20. Fig. 21. Fig. 22. PLATE 5. G. heteroclita. Dorsal view of a rather small individual. For clearness furrows between annuli are shown only at the margin of the body, except where they mark somite boundaries. Salivary glands are shown only in the right half of the figure, testes only in the left half. From an entire prepara- ration. XX 62. Combination of two or three successive parasagittal sections of head end, — X 83. Brain viewed from the left side. From several sections combined. Ventral view of a living animal bearing eggs. X about 13. ’ / ——)" a —— B Meisel, lith. Bost. Ee CasTLE. — Rhynchobdellide. PLATE 6. G. elongata. All figures of this plate were drawn from whole preparations. Annuli in Figures 23 and 27 are numbered from the posterior margin of the oral sucker backward. ~ Fig. 23. Fig. 24. Fig. 25. Fig. 26. Fig. 27. Head end viewed from right side. Posterior end of body viewed from right side. Brain, ventral view. Brain viewed from right side. Ventral view of entire animal partially contracted. In somites v11.—xx11. furrows between annuli are shown only at the margin of the body, except where they mark somite boundaries. PLATE 6. B. Meisel, lith. Bestia. CAsTLE, — Rhynchobdellide. Fig. 28. Fig. 29. Fig. 30. Fig. 31. PLATE 7. G. elegans. Dorsal view of a young individual. In somites v1.—xxy. furrows between annuli are shown only at the margin of the body, except where they mark somite boundaries. Reproductive organs and salivary glands drawn from other, older individuals; salivary gland cells a little too small. From an entire preparation. Parasagittal section of head end. Head end, dorsal view. From an entire preparation. X about 40. The same, ventral view. B Meisel, fith Bestes.. Fig. Fig. Fig. Fig. Fig. Fig. Fig. CasTLE. — Rhynchobdellide. 30. 34. 38. 36. 37. 38. PLATE 6&8. G. parasitica, var. plana. Dorsal view of head end of an individual from Havana, Illinois, in which the division between rings 2 and 3 was not evident. Stippling shows position of yellow pigment in a median vitta and (on left side) in metameric marginal spots. From an alcoholic specimen. X 41. ; G. parasitica, var. rugosa. Dorsal view of head end of an individual from Cambridge, Mass., showing the usual annulation of somites 1-111. From an alcoholic specimen. Enlarged. G. stagnalis. Dorsal view of posterior end of body. Enlarged. G. heteroclita. Dorsal view of head end of a living animal, showing most common position of eyes. Enlarged. P A dorsal view of the head end of the individual represented in Figure 388. The anterior ring of somite vi. is seen to contain traces of a trans- verse pigment line. Drawn from the living animal. Enlarged. G. parasitica, var. plana. Dorsal view of posterior end of body of the individual shown in Figure 32. Marginal light spots indicated by stippling. x 24. G. heteroclita. Dorsal view of a living animal, showing the general form of the body at rest, and the color pattern sometimes present on the dorsal surface. The rings are not indicated, but the numerals are placed opposite and serve to designate those rings in which the pigment is found (the anterior rings of their respective somites). Enlarged. 2], lith. Bastion 7 ES: DM — 5 = » - 62 Length of hypohyal asta tershr Steed aan wise oe op Wis) Mopamomecmiohyal . . . . . 1 ewe ww DB Length of epihyal Si SUP ce ae aR eeOm Er me bon 63 wane) 0) PIOUeNmECIMyIGlO evs ss fs ee 5 et es | OD HernuomeanGalpedicle 9 2 3 2. ee. te sw 6 Width of basioccipital concavity . . . . Ri tcmeisirs panes Distance from basioccipital concavity to vomer . . . . 16.0 . a ee ee ee eee eee astman—Lepidosteus. 4 E PEATE. 1- THE HELIOTYPE PRINTING CO., BOSTON. 2 PLATE astman—Lepidosteus. E BOSTON, THE HELIOTYPE PRINTING CO Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE. Vout. XXXVI. No. 4. CHARACTERS AND RELATIONS OF GALLINULOIDES, A FOSSIL GALLINACEOUS BIRD FROM THE GREEN RIVER SHALES OF WYOMING. By Freperic A. Lucas. Wirth One Puate. CAMBRIDGE, MASS., U.S. A.: PRINTED FOR THE MUSEUM. , AuvGust, 1900. No. 4.— Characters and Relations of Gallinuloides wyomingensis Eastman, a Fossil Gallinaceous Bird from the Green hover Shales of Wyoming. By Freperic A. Lucas. THE specimen upon which the following observations are based was discovered in the Green River Shales (Middle Eocene) of Fossil, Wyoming, during the summer of 1899, and was shortly after procured for the Museum of Comparative Zodlogy at Cambridge, where it is now preserved (Cat. Foss. Birds, No. 1598). Dr. C. R. Eastman briefly described (Geological Magazine, February, 1900) the bird as Gallinu- loides wyomingensis, and at his solicitation a more detailed investigation of its structure and relations was undertaken, the results of which are herein set forth. Like the well-known Green River fishes, the specimen is very complete and in a most excellent state of preservation, although a little injured as to skull, vertebree, and digits through the over-zealous preparation of the collector. There is a thin, dark, unctuous layer lying on the same plane as the skeleton and almost confluent with the thinner bones, so much so that in developing the finer points it was at times difficult to shun the temptation to carve out a character that might readily be imagined to exist. This layer obscures the ribs, which are scattered, as well as other portions of the skeleton. While, however, many structural details cannot be made out, the general characters are so distinct and the affinities of the bird so apparent that these defects are of compara- tively small importance. The Green River bird was of about the size of a Ruffed Grouse, but stood somewhat higher on its legs. Its galliform nature is obvious at a glance, the most apparent peculiarities being the length of the legs and the depth and the anterior extent of the sternal keel. The majority of its structural resemblances are with the curassows and with the genus Ortalis amongst those birds, but while according to Huxley’s definition it indisputably falls in the Peristeropodes, there are sufficiently strong characters to exclude it from both the Cracidze and Megapodiide. The bird presents no points of affinity with any of the American grouse, still less with any of the Odontophorine. VOL. XXXVI. — NO. 4. 80 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Cracine and Galline are herein used as short equivalents for “ peristero- podous” and “ alectoropodous,” — the latter terms, although expressing the precise meaning needed, being a trifle cumbersome for ordinary use ; “oalliform” is employed to designate such characters as are shared in common by all members of the Galliformes. Head. — The beak much resembles that of Ortalis, being moderate in size, stouter than in Crax, Rollulus, and Phasianus ; but not so short, stout, and decurved as in Colinus and allied genera. The holorhinal narial opening is also much like that of Ortalis, and the nasal, which has escaped injury, is typically galliform ; the superior process can be clearly seen, but the inferior process is covered on its lower part by crushed bone. The lachrymal, or prefrontal, appears to have been well developed, con- trasting in this respect with the American grouse (in which the prefron- tal is usually quite small), and agreeing with the curassows. The post- frontal process is stout and directed forwards. The mandible is stout and imperforate, and while it has a blunt angular projection, the re- curved process so characteristic of the Galliformes is lacking. This is the most notable departure from the galliform structure found in the skeleton. Vertebre and Ribs. — Little can be said of the vertebre save that the vertebral column presents the customary galliform arrangement of a free vertebra in front of the synsacrum preceded by a mass of anchylosed vertebre, but as to the number of the latter nothing can be affirmed. The cervicals have suffered from the mistaken zeal of the preparator, and but five can be definitely distinguished between what should be the axis and where the column disappears in the flattened bones of the wings. The caudals are mostly lacking, so that, unfortunately, nothing can be learned from them. Four pairs of ribs are articulated with the sternum, and at least one pair (one is the customary number in the Galliformes) arose from the synsacrum. Several ribs lie over the synsacrum, but there is no reason to suppose that all of them articulated with it. The usual number of ribs among the Galliformes is five on a side; Pavo has six, but the number in the present specimen cannot be made out. There is quite a little space between the first and second costal facets, the succeeding three being crowded together. This is interesting from the fact that it is a feature of modern galline birds, the spacing of the costals being more regular among the curassows. Shoulder Girdle. — The scapula is not unlike that of Rollulus, being long, narrow, and with parallel borders, as in many of the curassows, or LUCAS: GALLINULOIDES WYOMINGENSIS EASTMAN. 81 as in Pediocetes. The coracoid resembles that of the Old World pheas- ants, and especially that of Phasianus colchius, more than it does the corresponding bone in any of the curassows. The epicoracoid is a little " more angular than is customary among Galliformes, but the epicoracoid of Pediocetes is of much the same pattern, and in this small point the Green River bird makes its nearest approach to some of the American grouse. The precoracoid process appears to be absent, as it is in most Galliformes, although there is a suggestion of this process in Arboriphila. The scapular process is small. The distal end of the coracoid makes a more obtuse angle with the shaft than is usual even in galline birds, but in this respect it is very similar to Phasianus colchius. Scapula, coracoid, and furcula, natural size. The furcula is unusually short and stout for a gallinaceous bird, ex- ceeding in this respect any species with which it has been compared ; it is U-shaped rather than V-shaped, most nearly resembling Numida in this particular. There is a distinct though slight acrocoracoid process, so that the furcula did not merely rest against the inner side of the coracoid, but articulated with it, thus differing from all existing Galli- formes. The scapular ends of the furcula are hidden so that it cannot be positively stated whether or not they reached the scapula. The hypo- cleidium is large and triangular, contrasting with Crax, which has a spinous hypocleidium, and exceeding Ortalis, in which this process is subtriangular and of moderate size. The sternum has a manubrium of moderate size, but from the disposi- tion of the bones it is impossible to ascertain whether it is perforate or imperforate. Both the external clefts are quite deep, and the external as well as internal xiphoid process is directed well backward ; both 82 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. processes are expanded at the free end. In the specimens of curassows available for comparison the external xiphoid is not pedate, but there is a suggestion of this condition in Talegallus. The sternal clefts are typically cracine, there being no approach to the deep internal cleft which makes the external and internal xiphoids of galline birds really branches of one process. The keel of the sternum is produced more anteriorly than in other Galliformes, though nearly approached by Centrocercus. It is to be noted that in this latter form the furcula is unusually long and narrow. Fore-limb. — The humerus, like the other bones of the wing, is stout and has the deltoid process well developed. The crushing which the bone has undergone prevents its being definitely stated whether or not the humerus was pneumatic, although the probabilities are that it was not. The structure of the wing, in conjunction with that of the sternum, indicates a bird of good powers of flight. The other bones of the wing lie so nearly over one another and are so flattened together that little can be said as to their details, save that the third metacarpal appears to have been much straighter than is usual among gallinaceous birds. Pelvic Girdle. — As the pelvis lies on its dorsal surface it cannot be stated whether or not it was curved or straight in profile, but in the subequal proportions of the pre- and post-acetabular portions it resembles the curassows, although the conditions are much the same in Meleagris. It is somewhat wider in comparison with its length than in the curas- sows, the proportions resembling those observed in Thaumalea. There is no tendency toward separation of the ilia and ischia. The ischia do not seem to be bulged out to overhang the pubes as they do in Ortalis, but this feature is so extremely variable in the Galliformes as to have little or no significance. The pubes are long and slender, and as the speci- men now lies, they appear parallel with one another throughout their distal halves. In most Galliformes the pubes approach each other distally, sometimes, as in Ortalis and Penelope, being almost in contact. In this respect the Green River specimen departs from the cracine type and approaches such forms as Meleagris and Rollulus, and while it is of course possible that the pubes may have approached each other in the living bird, the intervening space is now so great as to make this seem doubtful. The prepubis is small, the obturator foramen very small, and the ilio-ischiadic space moderate. Hind-limb. — The femur is so crushed as to oben its characters. There is no sign of a patella, though this may have been present. The cnemial ridges are slight, and there is the customary osseous tendinal apy LUCAS: GALLINULOIDES WYOMINGENSIS EASTMAN. 83 bridge on the anterior face of the distal end of the tibia. The fibula is of the same general proportions as in other Galliformes. The hypotarsus is very likely only grooved, not perforate ; but this is one of the points that cannot be definitely ascertained without injury to thespecimen. The number of tarsal tendinal perforations is a character of much importance in birds, for it seems fairly constant within the limits of a given large group and indicates the amount of specialization attained by the members of that group. As all Galliformes examined have a single tendinal perforation, the absence of such a character would indicate that our Eocene bird is of a more primitive type than its modern relatives. The usual tarsal sesamoid shows back of the right tarsal joint. The tarsus is longer in proportion to the tibia than in any other species examined, as is shown by the subjoined table, which gives the length of these bones in a few species : — SPECIES. LENGTH OF TIBIA. LENGTH oF TARSUS. Ratio. Gallinuloides wyomingensis 57° mm. 45° mm. 1.27 Penelope superciliaris 115. 82. 1.40 Rollulus roulroul 72. 48. 1.50 Phasianus colchius 112. 72. 1.56 Ortalis maccalli 108. 65. 1.66 Colinus virginianus 53. 30. lec The toes are moderate and slender, of about the same length as those of Colinus virginianus, but a little heavier ; yet they are not heavy in comparison with the size of the tarsus or the general bulk of the bird. The following table gives the length of the principal bones in the skeleton, all measurements being made in a straight line : — PRINCIPAL MEASUREMENTS OF GALLINULOIDES WYOMINGENSIS. Occipital condyle to tip of bill, 47.°mm. Xiphoid to anterior end of keel, 59+ mm. Humerus, 47. Femur, 41. Ulna, 49 + Tibia, 57.+ Metacarpus, 25. Tarsus, 45. Scapula, 48. Basal phalanx of digit L., 7.5 Coracoid, 29. do do do ees ile Xiphoid to manubrium, 59. + do do do EE 1194. do do do Ves 7.5 Relationships. — The various characters of the Green River bird may be summarized as follows : — 84 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Galline Characters. — Pedate end of internal xiphoid process, arrange- ment of the costal facets, and shape of the distal end of coracoid. Cracine Characters. — Blunt, upright, subtriangular costal process, shallow inner sternal notch, small prepubis, proportions of pelvis, elon- gate tarsus with all the toes on the same level. Peculiar Characters. — Absence of recurved mandibular process; short, stout, U-shaped furcula with large hypocleidium and articular facet for coracoid. The weight of the peculiar characters, particularly the absence of a post-angular process, are, as stated in the introductory remarks, sufficient to prevent the bird being placed in either the Cracidee or Megapodide, thus necessitating the establishment of a new family, Gallinuloidide. The principal family characters are the absence of a postangular man- dibular process, presence of an articular facet on the furcula for the re- ception of the acrocoracoid, and the presence of an acrocoracoid. The generic characters are considered to be the stout U-shaped furcula, the shape of the scapula, and the anterior extent of the crista sternt. As specific characters are always comparative, none can be - formulated from a single specimen, even did they not depend to so great an extent in birds — often entirely — on external features. This bird is interesting not because it presents any striking peculiari-~ ties of structure, but rather because it does not, and because it belongs, as we might naturally expect from its age, to a generalized type having points of structural resemblance with various families of gallinaceous birds. It is an additional reminder, were any needed, of the great gaps in our knowledge of the development of birds and of the rapidity with which they attained their present forms. The mammals of the Eocene are quite different from existing species, but this bird readily takes its place among the forms of to-day. Lucas—Gallinuloides. PLATE 1. TE Fy OSE RH Seiad aI een “2 a aatte ne ARE FN jax” , THE HELIOTYPE PRINTING CO BOSTON, oe ee Bulletin of the Museum of Comparative Zodlogy — AT HARVARD COLLEGE. | Vou. XXXVI. No. 5. DEVELOPMENT OF THE MOUTH-PARTS OF ANURIDA MARITIMA GUER. ; By Justus Watson Fo.rsom. Wits Ereut Puates. CAMBRIDGE, MASS., U.S.A.: PRINTED FOR THE MUSEUM. OctoBErR, 1900. No. 5.— The Development of the Mouth-Parts of Anurida maritima Guer4 By Justus Watson FoLsom. CONTENTS. PAGE PAGE Introduction ...... . . £87|Linguaand Superlingue.... 110 Methods. . . eee Se axilla, ov te a le ee WALD General Reeeription of ee ee SO nab. oe os” 6: eae tes ts LO Meemeisiaees. . - . . - =. O0|Skull. . ~~. - + 2 ss = © 185 Procephalic Lobes . . . > . . 91|Tentorium. . . aS lay!) Labrum and Clypeus . . . . . 93} Segmentation of the Head oes om laz Antenne .. BL et = OG SMINEIAE 620 ce ct) ats, “sup Premandibular Appendage | 2.) Ge Biblioptaphy sce «es ROL Mandibles ... . >. «' 102) Explanation of Plates 2°. <3. Lot Introduction. Our present ideas of homology in the details of insect mouth-parts rest almost exclusively upon anatomical data, and need careful revision in the light of embryological facts. Too many entomologists have speculated upon the subject in complete disregard of evidence from ontogeny or phylogeny. Embryologists, on the other hand, have greatly neglected the mouth-parts. It seems almost superfluous to insist that highly specialized organs can be but imperfectly understood unless studied in egg and larva as well as imago ; that generalized types illuminate specialized forms ; and that equivalent groups are linked together through their more general- ized members ; yet too often these accepted principles are not applied. The objects of the present paper are two: first, to supplement my pre- vious account (Folsom, ’99) of the anatomy and functions of the mouth- parts of a representative Collembolan ; second, to discuss the morphology ‘of mandibulate mouth-parts of insects and their nearest allies upon anatomical and embryological evidence derived from the most primitive insects, the Apterygota. 1 Contributions from the Zodlogical Laboratory of the Museum of Comparative Zoviogy at Harvard College, E. L. Mark, Director, No. 114. VoL. xxxvi.— No. 5 88 BULLETIN: MUSEUM OF COMPARATIVE ZOSLOGY. My comparisons have been hindered by the scanty and fragmentary nature of published embryological observations upon the mouth-parts of Arthropods. Detailed studies upon the subject in the less specialized Pterygota, Crustacea, Arachnida, Diplopoda, and Chilopoda do not exist, but are necessary for the proper understanding of the morphology of the mouth-parts, and will have much bearing upon the phylogeny of the classes named. The present study was made under the supervision of Dr. C. B. Daven- port, to whom I am most grateful for his constant, critical Bia valuable advice and encouragement. Professor E. L. Mark has carefully revised the text and attended to all the details of publication ; his help, as always, has been of inestimable value to me. Methods. For killing eggs, and adults as well, simply hot water was used, with excellent results. After killing, material was carried through several successively stronger grades of alcohol and finally preserved in absolute alcohol. In the study of the embryo, both dissections and serial sections were made. As much as possible was learned by dissection, as that method, although difficult, gave more trustworthy results than could possibly be obtained by reconstruction from sections. The germ bands of freshly killed embryos were too delicate to be dissected out uninjured ; but after being in absolute alcohol for two months they had become sufficiently hardened for this operation. A longer stay made them brittle, but advantageously so in some respects. Dissections were made under a compound microscope with a magnifi- cation of about one hundred and fifty diameters. For the finest work, the “‘minutien Nadeln,” used by entomologists for pinning minute in- sects, were employed, The general form and position of an embryo could be seen through the transparent egg-membranes; but to get clearer views, the outer membrane was removed, the remaining corru- gated membrane punctured, and a staining fluid allowed to penetrate the germ band. Preparatory to the dissection of minute structures, the egg was placed in weak glycerine, which caused the embryo to shrink away from the membranes slightly, allowing these to be removed ; the germ band was dissected out and stained with Grenacher’s alcoholic borax-carmine or hematoxylin, Isolated parts of the embryo were mounted temporarily in weak glycerine without pressure, in such a FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 89 way that by moving the cover glass they could be rolled into various desirable positions. 5 For sectioning, portions of the embryo, or punctured eggs, were im- bedded in hard paraffine. The eggs required at least four hours for thorough penetration. For orienting, Woodworth’s (’93) method was employed, but not always with success, as the objects were liable to be- come distorted or even lost. A simpler, but more efficient, method for these particular objects was to orient them under the compound micro- scope with a hot needle in a glycerine-smeared watch-glass of melted paraffine, and to fix them in place by touching the glass beneath with cold water before hardening the paraffine throughout. When the block of paraffine was inverted under the compound microscope, the imbedded object could be seen through a thin film of paraffine, and a scratch could be made to indicate the plane of sectioning. Sections from 5 to 10 w in thickness were cut with either a Reichert or a Minot-Zimmermann microtome, fastened with Mayer’s albumen mixture, and stained with various reagents, chiefly Delafield’s or Klein- enberg’s hematoxylin followed by safranin, Grenacher’s alcoholic borax- carmine, and Heidenhain’s iron-hematoxylin. General Description of Egg. The eggs of Anurida maritima are spherical, from 0.26 mm. to 0.38 mm. in diameter, enlarging with age, and at first light yellow, later becoming orange. They occur abundantly along the Atlantic coast under stones between tide-marks, and are usually mingled with the conspicuous white exuvie of the parents. The eggs of Collembola depart widely from those of other insects by being holoblastic; they are slightly unequal in cleavage. After the mo- rula stage the outer nuclei and accompanying protoplasm migrate toward the periphery, leaving behind yolk masses and also cells which subse- quently prove to be entodermal. The peripheral cells become arranged in two layers: the ectoderm, a continuous superficial layer, with nuclei at regular intervals, and the mesoderm, an inner, less compact layer with fewer and scattered nuclei. Thus there soon results a condition like that derived from superficial cleavage. The ventral plate, or germ band, is formed by migrant mesoderm cells, and, according to Uzel (’98, p- 22, Tomocerus), is first represented by two pairs of isolated thicken- ings, —the procephalic and mandibular fundaments. I have found 90 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. that the appendages appear in succession from in front backward, and that they are well developed long before the segmentation of the germ band. The blastoderm is interrupted only by the “dorsal organ,” which is attached to the inner egg membrane. Claypole (98, pp. 255-258) distinguishes five egg membranes in Anurida, and maintains that all arise from the egg or the blastoderm. I find that in the ripe egg two are evident: a thick outer and a thin corrugated inner one, respectively analogous to, if not homologous with, the chorion and the vitelline membrane of other insects. Another deli- cate membrane completely envelops the embryo in early stages (Plate 1, Figures 1, 3, mb.), except where interrupted by the dorsal organ. I have found it to be, not a “ larval skin,” but a blastodermic membrane. The peculiar cleavage of Collembola has been observed by Oulganine (75, ’76), Lemoine (’83), Claypole (’98), and Uzel (’98). In the most nearly related group, Thysanura, the cleavage has been shown to be superficial by Grassi (85), Heymons (’96, ’97°), and Uzel (97, ’98). In cleavage, then, Collembola resemble many Crustacea and Arachnida, in which it is at first total and secondarily superficial. Thysanura, on the other hand, approach the Orthoptera, in that the cleavage is from the first superficial. The ‘dorsal’? or “ precephalic”” organ of Collembola has been de- scribed by Lemoine (’82), Wheeler (93), Claypole (’98), and Uzel (’97, 98); of Thysanura, by Grassi (85), Heymons (’96, ’97*), and Uzel (97,98). Wheeler homologized it with the “ indusium ” of Orthoptera, and suggested its analogy with the embryonic sucking-disk of Clepsine. Claypole collected evidence of a similar structure in Crustacea, which has been reinforced by Uzel. Reference Stages. For descriptive purposes I have selected nine consecutive stages of development, which may be identified in the entire egg by the following characteristics : — At Stage 1 (Plate 1, Figure 1) the embryo is almost spherical with all the primary appendages represented by small papillae. The dorsal organ is large, with a spherical imbedded portion and an expanded super- ficial part, the latter firmly attached to the corrugated membrane. This stage is very nearly that of Claypole’s (’98, Plate XXIII.) Figures 40 and 41 of the same species. Stage 2 (Plate 1, Figure 2) is characterized by folds representing the FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 91 last five abdominal segments, and by longer appendages, of which the antennz and legs show traces of segmentation. It is approximately the stage of Figures 42 and 47 of Claypole. At Stage 3 (Plate 1, Figure 3) the ventral surface of the embryo is almost flat, preparatory to involution ; the legs are decidedly longer, and the fundament of the proctodzum is distinct. Figures 43 and 432 of Claypole belong near this stage ; also Figure 10 of Ryder (86), likewise for Anurida maritima. During Stage 4 (Plate 1, Figure 4) the germ band is folding into the yolk, the fold beginning anteriorly and continuing backward. The antenne and legs are long and stout. My figure shows a stage a little later than that of Figure 44 by Claypole. At Stage 5 (Plate 1; Figure 5) the involution has reached the centre of the egg, the antenne and legs are distinctly segmented, the mouth- folds are conspicuous, and the dorsal organ has shrunken considerably. Stage 6 (Plate 1, Figure 6) is much like the last, except that the head and tail of the embryo have approached each other. The dorsal organ is much reduced and somewhat flask-shaped. This is the stage of Ryder’s Figure 7. At Stage 7 (Plate 2, Figure 7) the eyes are first recognizable as five black circular patches on either side. Figure 45 of Claypole represents this condition. Stage 8, which I have not figured, differs externally from the last in that the number of eyes is no longer evident, it being obscured by a suffusion of pigment. The degenerating dorsal organ now disappears by resorption. Stage 9 (Plate 6, Figure 41; also Claypole, Figure 48) refers to the newly hatched insect. Before this period, movements of the insect may be seen through the egg membranes. If eggs have been kept dark, — the normal condition, — the emerging insects are white, excepting the eyes ; if exposed to sunlight, however, the embryos become blackish- blue long before hatching. At emergence the external clothing of setz is complete, and the mouth-parts are fully formed. Procephalic Lobes. The fundaments of the procephalic lobes are two isolated thickenings of the blastoderm, which are the first of the paired fundaments to ap- pear. Each procephalic fundament is lenticular in form and rapidly increases in thickness and area. In the earlier stages the procephalic 92 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. thickenings are not definitely circumscribed, but merge insensibly with the rest of the blastoderm. Previous to Stage 1 the procephalic lobes meet in the median plane, where the labral fundament then appears. Before the appearance of the labrum, however, the antennal fundaments evaginate from the pos- terior regions of the procephalic lobes. In Stages 1 and 2 (Plate 1, Figures 1 and 2) the lobes continue to increase in area and thickness. . At Stage 3 either lobe is relatively as thick as is represented in Figure 3, pr’ceb., and in lateral surface views (Plate 2, F =< 9, 10, pr’ceb.) appears as a strongly convex, oval protuberance. In Stages 4 (Plate 3, Figure 12, pr’ceb.) and 5 (Plate 3, Figures 19, 20, 21, pr’ceb.) the procephalic lobes change little except in size, and the median depression between them (sz/.) is still distinct. In Stage 7 (Plate 5, Figure 30) the depression (su/.) becomes obliter- ated, and the eyes (Plate 4, Figure 24, ocl.) and postantennal organs (Figure 24, o.p’at.) appear. At this stage sections show a pair of gan- glionic fundaments (Plate 4, Figure 28, pr’ceb.), the largest and most anterior in the head, with which the next two pairs eventually unite to form the supra-cesophageal ganglion of the adult (Plate 8, Figure 51, gn.swe@.). Tn other Collembola the procephalic lobes develop in just the same way, as may be gathered from Nicolet (’42, Smynthurus), Packard (71, Isotoma), Lemoine (’83, Smynthurus), and Uzel (98, Tomocerus). Also in Campodea the same course of development is followed (Uzel, ’98, Taf. 3, Figures 33-36; Taf. 4, Figures 37-42) as well as in Lepis- ma (Heymons, ’97*). In fact, the simple process described for Anurida characterizes not only Orthoptera (Ayers, ’84, Wheeler, ’93, Heymons, ’95), but also in- sects in general. The procephalic lobes of Diplopods and Chilopods develop essentially as in insects and Crustacea, but no detailed comparisons can be made as yet. The most interesting considerations concerning the ocular segment of Hexapoda relate to its equivalence with the first segment of Crustacea. Viallanes (’87, pp. 98-109) has carefully compared the brain in both classes and found a striking agreement, extending to histological details : ‘‘Considérons en premier lieu la partie latérale du protocérébron, connue des anatomistes sous le nom de ganglion optique ; elle nous montre d’abord, en allant de dehors en dedans, les parties suivantes: les fibres FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 93 post-rétinennes, la lame ganglionnaire, le chiasma externe, la masse médullaire externe, le chiasma interne et la masse médullaire interne. “Toutes ces parties, si nettement caractérisées, se retrouvent sans modification chez l’Insecte; il n’est done pas douteux qu’il existe au moins pour cette premiére région du ganglion optique similitude com- pléte entre les deux types que nous cherchons 4 comparer. . . . Cette similitude a été reconnue par tous ceux qui se sont occupés de ce sujet (Berger, Bellonci, Carriére et moi). . . . En somme, au point de vue des parties dont nous venons de parler, il n’existe que des différences bien peu importantes entre I’Insecte et le Crustacé: chez le premier, les deux lobes cérébraux sont tres rapprochés et se soudent sur la ligne médiane ; chez le second, ces mémes parties (appelées balles supérieures) sont écartées, chacune d’elles étant logée dans le pédoncule oculifeére correspondant.” Packard ’98, p. 51) says, “Hence the ocular segment, 7. ¢., that bearing the compound and simple eyes, is supposed to represent the first segment of the head. This, however, does not involve the conclusion that the eyes are the homologues of the limbs, however it may be in the Crustacea.”’ As Viallanes has proved the equivalence of protocere- brum and optic nerves in insects with those of Crustacea, and others have shown that the compound eyes of both groups are constructed alike, even to the number of retinal elements, it is proper to infer that the compound eyes of the two groups are homologous. The protocerebrum of Collembola and Thysanura agrees in develop- ment and structure with that of other insects and also with decapod Crustacea ; the facetted eyes of Hexapoda and Crustacea are likewise homologous. Labrum and Clypeus. The labrum is chiefly interesting because it has frequently been held to represent a pair of primary appendages. At Stage 1 (Plate 1, Figure 1; Plate 2, Figures 8, 8*, ddr.) the labrum (really clypeo-labrum) is a median hemispherical papilla anterior to and distant from the bases of the antenne ; at no period does it give evidence of a paired origin. At Stage 2 (Figure 2), while the distances between the labrum and mandibles is precisely the same as in the preceding stages, the antennz are inserted beside the oral region of the upper lip; the latter is globular and flattened against the egg shell. 94 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Surface views at Stage 3 (Figure 3) are given in Figures 9, 10, and LI, lor. A sagittal section at this stage shows (Plate 3, Figure 13) an elon- gation of the labral fundament, and demonstrates its origin from the germ band by simple evagination. The posterior surface of the labrum is now the anterior wall of a distinct invagination (07.), the fundament of the stomodzum. At Stage 4 (Figure 4) the labrum is longer (Plate 3, Figure 19, br.) and its long axis has swung backward, probably on account of the ex- cessive elongation of the anterior labral surface. A ventral aspect of the germ band (Figure 12) shows the labrum to be approximately oval in cross-section, but with a more rounded anterior surface. At Stage 5 (Figure 5) the labrum (Plate 3, Figure 20, Jbr.) is de- cidedly longer. The basal part of the labral fundament represents the clypeus, with which the lateral folds, or mouth-folds (Figure 21, pli.or.) are now confluent ; overhung by the end of the labrum is the distinct stomodzum. At Stage 7 (Figure 7) a distinct depression (Plate 5, Figure 31, dep.) separates the clypeus from the procephalic lobes ; the depression, in fact, may be seen as early as Stage 1, for it simply forms the angle be- tween the labral fundament and the procephalic lobes. Although the clypeus merges insensibly into the cheeks, the labrum is a free trapezoi- dal plate, as in the adult (Plate 6, Figure 40, /br.). The antenne are now inserted (Plate 4, Figure 24, Plate 5, Figure 30, at.) almost ex- actly opposite the base of the labrum. At this stage the clypeo-labral suture is not distinctly indicated (Figure 24), but in Stage 8 an invagi- nation occurs to form the labral hinge of the adult (Plate 6, Figure 40, ate.). In Stage 8 the only other important changes in the labrum are the evagination of single hypodermis cells to form the external sete, and the formation of trivial cuticular folds which represent the rudiments of the epipharynx. In Anurida, as in Orchesella, the epipharynx is purely a cuticular structure and unconnected with the central nervous system. In the adult Anurida a shallow clypeo-frontal groove is distinguish- able (Plate 6, Figure 40, su/.), but does not amount to a suture, and the clypeus is not laterally demarcated from the groove. In Orchesella and Tomocerus, however, the clypeus is a distinct sclerite. In none of the Collembola that I have studied is there any distinction between clypeus and labrum on the roof of the pharynx. gr” FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 95 Packard (’71, p. 18) says, regarding Isotoma, ‘‘ The clypeus, however, is merged with the epicranium, and the usual suture between them does not appear distinctly in after life, though its place is seen in Figure 13 to be indicated by a slight indentation. The labrum is distinctly defined by a well-marked suture, and forms a squarish knot-like protuberance, and in size is quite large compared to the clypeus. From this time begins the process of degradation, when the insect assumes its Thysanurous characters, which consist in an approach to the form of the Myriapodous head, the front, or clypeal region being reduced to a minimum, and the antenne and eyes brought in closer proximity to the mouth than in other insects.” Lemoine (’83, p. 510, Planche XV., Figure 24) mentions in Smynthu- rus, “Les deux appendices qui constitueront la levre supérieure,” but they appear in his figure as only simple lobes from a large, median labrum. Wheeler (’93, p. 57, Figure VI.) represents the labrum of Anurida as a median, unpaired fundament, and Claypole (’98, Plate XXIII.) gives several surface views of the upper lip in the same species. Uzel (98, Taf. VI., Figur 87) shows the single labral fundament of Macrotoma (Tomocerus). Regarding Campodea, Uzel (98, p. 26) says: “ Vor der Mundeinsen- kung erblickt man jetzt schon die unpaare Anlage der Oberlippe,” and partially illustrates (Taf. IV., VI.) the development, which proceeds essentially as in Anurida. The finished labra of Campodea (Grassi, 86°, Tav. IV., Figura 7) and Japyx (Grassi, 86", Tav. II., Figura 15 bis) are very simple rounded plates. For Lepisma, Heymons (’97*, Taf. XXX.) figures the labral funda- ment as a prolongation from the procephalic lobes, and characterizes it (p. 591) as “eine kleine, vollkommen, ungetheilte, einfache Platte.” Later (p. 593) he says, “ Die Oberlippe wird bedeutend grésser und bekommt an ihrem hinteren Rande eine mediane Einkerbung (Figure 17).” The median indentation is clearly, however, a secondary formation. In both Lepisma and Machilis (Oudemans, ’88, Taf. I., Figur 3) the labrum remains simply an anteriorly rounded plate. In the Orthopteran Cicanthus, Ayers (’84, p. 240, Plate 18, Figures 21, 22) describes the unpaired fundament which forms the ovate labrum. In short, the labrum in all Orthopteran families develops from an un- paired fundament. (See Wheeler, ’93, Heymons, ’95°.) The same is true of the Libellulide and Ephemeride (Heymons, ’96, 96 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Taf. II., Figuren 19, 29), and examples might be multiplied to show that the labrum does not represent a pair of appendages. The view held by Kowalevsky, Carriere and others, that it did, was based chiefly upon anatomical evidence, which has since been disproved by Heymons (95) and others. (See Packard, ’98, pp. 42-43.) Scolopendrella (Latzel, 84, p. 8, Taf. 1, Figur 43; Grassi, ’86*, p. 15, Tav. II., Figura 6) has a simple, emarginate, six-toothed labrum, and, like Hexapoda, a distinct, subtriangular clypeus. Moreover, as Packard (98, p. 22) has affirmed, it has a V-shaped tergal suture, which exists also in the more generalized insects, but is absent in Myriopods. In Diplopoda, an upper lip is present as a transverse plate, fused, however, with the cranium. In Chilopoda, a similar labrum is present, but is not always basally fused, and frequently consists of three transversely placed sclerites. It originates as a simple median lobe (Heymons, ’97°, p. 4, Figur 1, Scolo- pendra). In Crustacea the upper lip is derived from a median, unpaired evagi- nation corresponding almost exactly in position with the labral funda- ment among insects. Among insects, then, the labrum and clypeus develop from a median evagination between the procephalic lobes, and give no satisfactory evi- dence of paired origin. The same statement applies also to Crustacea, and, as far as is known, to Myriopods. Antenne. The antenne are the first paired organs to appear. They develop from the posterior boundaries of the procephalic lobes, and at Stage 1 (Plate 1, Figure 1, Plate 2, Figure 8, at.) are stout cylindrical papille already faintly constricted into two segments. As Figure 8 shows, they are more lateral than the other paired fundaments, and at first far be- hind the labrum. Sections prove them to be simple ectodermal evagina- tions, like all the other appendicular fundaments. At Stages 2 and 3 (Plate 1, Figures 2, 3, at.) the antenne are longer and usually composed of three segments. In Figure 2 the fourth seg- ment, which normally appears later than Stage 2, is suggested. They have now moved forward to positions near the labrum; in Stage 4 (Plate 1, Figure 4; Plate 3, Figure 12, at.) they lie on the two sides of that appendage, and in Stage 5 (Plate 1, Figure 5; Plate 3, Figure 21, at.) they have attained a position farther forward than the upper lip, FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. oT In Stage 6 (Plate 1, Figure 6, at.) there is clearly indicated a fourth antennal segment, which in Stage 7 (Plate 2, Figure 7; Plate 4, Figure 24, at.) becomes more distinct. At this time the antenne are long and stout, and occupy a position still farther forward than before. At hatching (Plate 6, Figure 41) they are pre-oral, more slender, dis- tinctly segmented, and clothed with sete. Elongation of the antenne occurs throughout their entire length, judging from the number of cells in longitudinal alignment on the same segment at different stages of growth, and also from the frequency of karyokinesis in different parts of the appendage. Growth is more rapid, however, in the apical region, from which the segments are successively constricted. In all the oral fundaments, in fact, growth was inferred to be most rapid at the apex, although likewise occurring throughout the rest of the ectodermal layer. At the apex itself — and these remarks apply equally well to the legs —the hypodermal cells are larger and more turgid than elsewhere, projecting as minute lobes from the surface. The chromosomes are very small, but frequently so arranged as strongly to suggest mitotic division. At Stage 5 (Plate 4, Figure 28, dew’ceb.) an antennary ganglion sup- plying the antennal nerves, becomes evident, but finally fuses with the first and third ganglia, between which it lies, to form the supracesophageal ganglionic mass. In Thysanura the antennz develop essentially as I have described for Collembola, being likewise at first post-oral and subsequently pre-oral, as Uzel (98) has shown for Campodea and Heymons (’97*) for Lepisma. Such a migration of the antennz is, however, not peculiar to Aptery- gota, but is characteristic of all insects, Among Diplopoda but a single pair of antennal fundaments occurs (Heymons, ’97°, p. 7, Figur 2, Glomeris). Judging from their position in relation to the mouth, they are equivalent to the antenne of Chilo- poda, among which Heymons (’97°, p. 4, Figur 1, Scolopendra) has discovered two pairs of antennal fundaments. The pre-antennal rudi- ments in Chilopoda appear to represent the antenne of insects and the antennules of Crustacea, the second pair to be equivalent to the inter- calary appendages of insects and the antennz of Diplopods and of Crustacea. It can scarcely be doubted, in view of the researches of Viallanes (’87), that the antennee of insects are homologous with the antennules of Crus- tacea. In the author’s words (’87, p. 105): “Voyons maintenant le deuxiéme renflement cérébral du Crustacé décapode. Il est formé d’une 98 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. paire de masses nerveuses ventrales connues sous le nom de lobes olfacti/s, réunies lune a l’autre par une commissure transverse, et d’une paire de masses dorsales qu’on pourrait désigner sous le nom de lobes dorsaux. ‘Les lobes olfactifs ont une structure tout a fait spéciale; la sub- stance ponctuée qui entre dans leur constitution est, pour ainsi dire, ‘segmentée’ en un grand nombre de petites boules d’aspect absolument caractéristique, qu’on désigne sous le nom de glomérules olfactifs. Les lobes dorsaux, au contraire, n’ont dans leur structure rien qui soit spécifique. “Le nerf antennaire nait du deuxiéme renflement cérébral par deux racines, — l’une sort du lobe olfactif, autre du lobe dorsal ; ce dernier, en outre, donne naissance a un nerf tégumentaire. “Cette description du deuxiéme segment cérébral du Crustacé peut, et sans qu'il y ait aucun changement a y faire, s’appliquer a I’ Insecte, tant il y a au point de vue de cette région cérébrale similitude entre les deux types. Nous sommes done en droit d’exprimer cette similitude, en appelant du méme nom de deutocérébron le deuxieme segment cérébral, qwil s’agisse @un Crustacé on d’un Insecte.” In favor of contrary views little can be said. ‘‘ Arguments drawn from the absence or presence of either pair of antenne in the higher Crustacea are not convincing, as there is great variation in the degree of development of their appendages in different groups” (Claypole, ’9g, pp. 265-266). Thus, in some Amphipods, the antennules are short, and in certain Isopods, extremely reduced. On the other hand, as Claypole notes, it is suggestive that in the generalized genus Apus, the first antenne are constant and the second variable or absent. The fact that the antennules of decapod Crustacea cannot be called “ post-oral” in origin, is not as significant as it may appear to be. The antennules originate at the side of the labrum (Reichenbach, g6), nearly post-orally, and migrate forward. In view of all other fun- damental correspondences between hexapod antenne and crustacean antennules, the trifling difference in original position may be ignored, especially as the organs in question are eminently migratory. I believe, therefore, that the deutocerebrum of Apterygota, repre- senting the second somite, is homologous with that of Orthoptera and other insects. Premandibular Appendages. The little-known “ premandibular” or “ intercalary” appendages are important as bearing upon the larger and much-disputed question of the segmentation of the head. FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 99 In Anurida, they are visible in Stages 1 and 2 only, as slight thick- enings of the germ band, which are often ill defined in outline and hardly deserve the name of appendages. In fact, their demonstration is largely a matter of technique. I dissected over thirty germ bands for this purpose, stained them variously, and mounted them temporarily in weak glycerine, without finding more than suggestions of the inter- calary appendages. At this point, Miss Claypole most kindly sent me some preparations which were a little clearer than any I had made. These I imitated by staining with Delafield’s hematoxylin, decolorizing with acid alcohol and mounting withont pressure in xylol balsam. If care is taken in decolorizing, a condition may be obtained in which all of the germ band between the antenne and mandibles has lost color excepting a rather vague patch on either side, usually not as distinct as in Plate 2, Figure 8*, app. pr’md. These patches are so slightly, if at all, elevated that they are not distinguishable with certainty in transverse or sagittal sections of the germ band. In good preparations, the lateral boundary of either appendage is indicated by a curving row of ectoder- mal nuclei, and this resemblance to the other paired fundaments is further shown in the presence of an imperfectly developed core of meso- dermal nuclei (Figure 8*, ms’drm.). Wheeler and Claypole have represented the appendages much smaller than I have, and appear to have figured the mesodermal core only. In none of Miss Claypole’s slides are the appendages outlined as sharply as in the preparation from which my Figure 8* was made. In glycerine the yolk granules interfere with proper observation, but in balsam this disadvantage is removed. Although the appendages are extremely rudimentary, the evidence they furnish of the presence of an intercalary segment is reinforced by the condition of the nervous system, for there is at Stage 5 a small neuromere (Plate 4, Figure 28, érv’ceb.), which, from its relation to the remaining cephalic neuromeres, must be regarded as belonging to the premandibular segment. It ultimately fuses with the deutocerebrum to form a part of the supracesophageal ganglion. Viallanes first called attention to the tritocerebral segment of insects and Crustacea ; he was aftérwards supported by Wheeler, who found that it bore a pair of appendages in Anurida; thus Wheeler (’93, p. 57, Figure VI.) discovered the intercalary appendages in this species, and indicated their obscurity by representing them by broken circles. Claypole (98, p. 263, Plate XXIII, Figures 40, 47) also observed the appendages, but erroneously inferred that they became modified 100 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. to form the sides of. the face,—a view which I shall discuss presently. A somewhat similar pair of appendages in the embryo of Apis was long ago observed by Biitschli (70), and a few years later by Grassi (85) also; but Packard (’98, p. 52, Figure 35) questions whether these belong to the category of segmental appendages. Heymons (’95°, Taf. I., Figur 5) also has recognized the “ Vorkie- fersegment ”’ in Orthoptera. He says (’95*, p. 16): “ Letzteres [Vor- kiefersegment] kommt, wie schon gesagt, tiberhaupt nur in ganz rudimentirer Weise zur Anlage. Extremitaten treten an ihm nicht mehr auf. Sein Ganglion riickt nach vorn und geht in die Formation des Gehirns ein. Bei dieser Gelegenheit werden zugleich auch die fus- serlich wahrnehmbaren Spuren des Vorkiefersegments verwischt. Selbst im Innen liegen die Verhaltnisse nicht viel giinstiger. Das Mesoderm des Vorkiefersegments bildet niimlich bei den Orthopteren ein eigenar- tiges Organ, den sogennantes Suboesophagealkorper, welches ebenfalls nur eine provisorische Bedeutung besitzt und spiiter zu Grunde geht.” The same author (97%, p. 590, Figur II.; Taf. XXX., Figuren 17, 20), referring to the embryo of Lepisma, writes (p. 591), “‘Genau auf der Grenze zwischen dem verbreiterten vorderen Kopfabschnitt und dem darauf folgenden verjiingten Kérpertheil zeigen sich ferner zwei, aller- dings nur schwach markirte, laterale Vordickungen (Z7c.). Dieselben kennzeichnen die Region des rudimentiiren Vorkiefer- (Intercalar) Seg- mentes. An diesen Segmente kommen wihrend der Entwicklung von Lepisma Extremititen nicht zur Ausbildung.” This nearly agrees with the condition in Anurida. Uzel records distinct intercalary appendages for Campodea in his pre- liminary paper (97°, p. 232), and in his final work (98, p. 26) says: “Sehen wir auf dem sogenannten Intercalarsegmente (Vorkieferseg- mente), das sehr deutlich entwickelt ist, jederseits eine kleine ErhOhung auftreten (int.), welche als die Extremititenanlagen dieses Segmentes zu deuten sind.” (p. 37.) ‘“ Die Extremititen dieses Segmentes werden bei Campodea in Form zweier Hocker angelegt, welche sich, wie wir voraussenden wollen, bis in das geschlechtereife Alter erhalten, und hier als Bestandtheile der ausgebildeten Mundwerkzeuge fungieren (der einzige bekannte Fall unter dem Insecten), indem aus ihnen kleine, praeoral gelegene, beiderseits an der Wurzel der Oberlippe befindliche Lappen (die Intercalarlappen, Taf. VI., Fig. 85, ant.) entstehen. Bei Lepisma sind keine Extremititenanlagen auf dem Intercalarsegmente vorhanden.” . . . “Unter den Myriopoden wurden von Zograf bei FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 101 den Embryonen von Geophilus ziemlich weit hinter dem Munde und’ dicht vor den Anlagen der Mandibeln zwei ansehnliche Hocker be- schrieben und abgebildet, welche wahrscheinlich den Hoéckern auf dem Intercalarsegmente von Campodea homolog sind. Sie werden nach dem erwahnten Autor immer kleiner und kleiner und sollen endlich ganz verschwinden.” In Anurida the intercalary thickenings become involved in the folds which form the sides of the face, as I shall describe, but I believe they are not, as Miss Claypole held, the fundameuts of those folds. In Tomocerus and Orchesella (Folsom, 99, p. 14, Plate 2, Figure 9) I have found that “at either end of the [labral] hinge. . . the cuticula is swollen into a conspicuous chitinous lobe, which projects into the pharynx to fit against a corresponding prominence of the mandible,” ete. As these lobes in the adult occupy precisely the same positions as those of Campodea (Uzel, ’98, Taf. VI., Figur 85, int.), I believe them to be intercalary appendages. In Anurida no such lobes exist. In Chilopods, two pairs of antennal fundaments appear (Heymons, ’97°, p- 4, Figur 1, Scolopendra), and the second, which alone become func- tional, are equivalent in position to the intercalary appendages of Apterygota as well as the antennz of Diplopods (cf. Heymons, 97°, p. 7, Figur 2, Glomeris). The equivalence of the tritocerebrum in Hexapoda and Crustacea was first shown in detail by Viallanes. His account (’87, pp. 105-108) is too long to be quoted in full, but he concludes: ‘ Les deux lobes con- stitutifs du tritocérébron de |’Insecte, et que j’ai désignés sous le nom de lobes tritocérébraux, représentent exactement les deux ganglions cesophagiens du Crustacé; ils donnent naissance aux mémes racines nerveuses, ils sont, comme ces derniers, unis au-dessous de l’cesophage par la commissure transverse de ]’anneau cesophagien.” Many authors (Korschelt und Heider, ’90-93, p. 906) agree in homolo- gizing the antennze of Hexapoda, innervated from the deutocerebrum, with the first antenne of Crustacea; also in homologizing the mandibles of both groups. Therefore only the intervening appendages of the trito- cerebrum remain to represent the second antennez of Crustacea. An intercalary segment, then, is to be recognized among Pterygota, at least in the more generalized forms, and especially among the primi- tive Apterygota, and in the latter group it may bear rudimentary appen- dages, even in the adult. The intercalary segment is to he regarded as equivalent in morphological value to any primary head-segment, — es- pecially because it bears a primitive ganglion, — and it constitutes the VOL. XXXVI.— NO. 5 2 102 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. third head somite. The tritocerebrum of Hexapoda is equivalent to that of decapod Crustacea, and the intercalary appendages of the former are homologous with the second antennez of the. latter, and probably with the antennz of Chilopoda and Diplopoda. Mandibles. The fundaments of the mandibles appear in Stage 1 (Plate 1, Figure 1, md.; Plate 2, Figure 8) as a pair of sub-hemispherical papille be- hind the antenne, and considerably nearer than they to the median plane. At Stage 2 (Plate 1, Figure 2, md.) they are longer and bluntly conical ; but at Stage 3 (Plate 1, Figure 3; Plate 2, Figure 9, md.) in lateral aspect they appear shorter than before, because the base is cov- ered by a lateral fold of the germ band (Figure 9, pli. or.). Sections through the mandibles transverse to the germ band (Plate 3, Figure 16) show that they are low broad ectodermal evaginations containing meso- derm. In Stage 4 (Plate 1, Figure 4; Plate 3, Figure 19, md.) the mandibles, although they have become long and cylindrical, are largely covered by the lateral folds ( pli. or.) which have grown more rapidly than they ; and in the following stage (Plate 1, Figure 5; Plate 3, Fig- ure 20, md.), though still nearly perpendicular to the germ band, they are almost completely covered laterally by the folds. The mutual rela- tions of mandibles and folds are shown in transections of the germ band (Plate 4, Figure 23), in which it may also be seen that the man- dibles (md.) are swollen at their ends, their lateral surfaces conforming to the adjacent surfaces of the folds (pli. or.). The long axes of the mandibles converge at their bases toward the median plane, and it is noteworthy that the lateral surface of each mandible is distinctly longer than the mesal surface (Figure 23, md.)—a foreshadowing of the oblique orifice of the finished organ. At Stage 7 (Plate 2, Figure 7; Plate 4, Figure 24, md.) the mandibles, now wholly covered by the lateral folds (plz. ov.), are much longer and still conical ; they are shorter and much more slender than the under- lying first maxille; and instead of being perpendicular to the germ band, they have now swung forward through an angle of almost ninety degrees; moreover, they converge in front toward the median plane, as do the first maxille (Plate 5, Figure 29). In this stage the mandibular muscles are individually distinguishable (Figure 32), and the anterior extremity of the mandible bears several minute lobes (Figure 32), each consisting of a single hypodermal cell. Inthe next (8th) stage the free a FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 103 end continues to bend toward the median plane until the apices of both mandibles meet. The terminal unicellular lobes become multicellular and secrete the incisive teeth (Plate 6, Figure 37 de. 7’cis.), of which there are finally five principal ones on the right and six on the left mandible. Although the “head” of the completed organ is almost solid chitin (Plate 6, Figure 37), there are five canals, one penetrating the base of each tooth ; the hypodermal cells have, however, receded from the “ head.” The extreme basal end of the finished mandible is prolonged as a chitinous, conical projection (Plate 6, Figure 36, cdz.), which, as in Orchesella, is let into a concave chitinous piece that I have called the stirrup (sta.), from which it may be withdrawn when the mandibles are protruded. This projection, or pivot, arises in Stage 7 (Plate 5, Figure 32, edz.) asa hypodermal evagiuation of the mandibular fundament, and simultaneously the chitinous stirrup (sta.) is formed in a transverse, superficial groove of the hypodermis lining the pharyngeal pocket in which the mandible lies. In Orchesella the lateral end of the stirrup unites with the external cuticula of the skull after traversing two layers of hypodermis: first, the layer lining the mandibular pocket, and second, the superficial layer of the head ; in Anurida, however, I have found no such union between stirrup and skull. The body of the man- dible is simply a modified cone, and hence in sections across this region appears as a complete chitinous ring (Plate 7, Figures 44, 45, md.). In Anurida no trace of a mandibular palpus exists at any stage, and, unlike Orchesella, no molar surface is differentiated ; the latter fact is correlated with the character of the food: Orchesella feeds upon ligni- fied vegetable substances, Anurida upon the soft tissues of the mollusk Littorina littoria. In further correlation with diet, the powerful rota- tors, or grinding muscles, of Orchesella are not represented in Anurida. Several writers on Collembola have already given surface views of the mandibular fundaments at early stages, although none have traced their development. I refer especially to Lemoine (83, Smynthurus) and Wheeler (93, p. 57, Figure VI., Anurida). Packard (’71, p. 17; Plate 3, Figure 13) evidently overlooked the mandibular fundaments of Isotoma, and what he regarded as mandibles are clearly, from their position, the first maxille. Ryder (’86) made the same mistake. Claypole (’98, Plate XXIII.) gives several figures of the mandibular fundaments of Anurida maritima before much differentiation has oc- curred, and Uzel (’98, Taf. VI., Figur 87) represents the fundaments in Macrotoma (Tomocerus) at a stage equivalent to that of my Figure 104 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 21. He also gives a figure (Taf. V., Figur 64) supporting his state- ment (p. 22) concerning the appearance of the mandibular segments: “ Ausserdem [collection of entoderm cells, etc.] bemerkt man zwei Paar . dunkler Stelle, welche an den Ecken eines gedachten Quadrats sich befinden. Das dem Dorsalorgane geniherte Paar dieser Blastodermver- dickungen (4/.) sind die getrennten Anlagen der Kopflappen, das zweite Paar (mds.) stellt die getrennten Anlagen des Mandibularsegmentes vor.” The eggs of Anurida at my disposal were either too old or too young to show the condition here described by Uzel, although I did find a stage in which three pairs of fundaments were present, the third pair being the first maxille. The mandibles probably follow the procephalic lobes in appearance, as I have found all the stages necessary to indicate that the remaining paired appendages, except those of the superlinguz, as I shall term them, appear successively from in front backward. Campodea is structurally nearest to the Collembola, and, thanks to Uzel (’98, Taf. III., Figuren 35, 36; Taf. VI, Figuren 77-85), some- thing is known concerning the development of its mouth-parts. The mandibular fundaments of Campodea are simple papille, as in Collem- bola ; this simplicity distinguishes the Apterygota from the most gener- alized Pterygota, the Orthoptera, in which the fundaments are sometimes lobed. The finished mandible of Campodea is strikingly like that of the Collembola, and is, moreover, of great morphological interest, because the structural correspondence of the mandible with the maxilla of hexa- pods — obscure in almost all other insects —is here a matter of direct observation, not merely one of inference. The mandible of Campodea (Meinert, ’65, Taf. XIV., Figuren 15, 16; Nassonow, ’87, p. 33, Figur 27) consists of a hollow fulerum (stipes) and a head, which is separated from the fulcrum by a transverse suture. The head is composed of two parts, — & large, toothed, immovable, outer lobe or galea, and a smaller, fringed, movable, inner lobe, representing the lacinia. Accepting the homologies with the first maxille implied in these terms, the palpus remains to be accounted for. A mandibular palpus has never been found among adult insects, — the evidence given for one by Hollis (’72) being quite vague and inadequate. Although the de- tailed development of the mouth-parts of Campodea has never been followed, it is in this most generalized insect that one may most hope- fully look for a trace of a mandibular palpus, and we may safely predict that, if found, it will be a lateral, distal lobe of the stipal region, just as it is in the maxillz of all insects. FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 105 The agreement between the finished mandibles of Campodea and Japyx, on the one hand, and Coliembola, as represented by Anurida and Orchesella, on the other hand, is remarkably close. In both groups the mandible is hollow, has an oblique basal opening, which is large in Cam- podea, and, instead of an ordinary articulation, a free basal pivot, which is peculiar to the Apterygota. The homologies extend further, for I find that the similar and complicated movements of the mandibles are actually effected by muscles which are probably homologous in the two groups. The equivalence of certain muscles in Campodea, as repre- sented by Meinert (65, Taf. XIV., Figure 15) with others figured by myself for Orchesella (Folsom, ’99, Plate 2, Figures 14, 15) may be ex- pressed in tabular form as follows : — Campodea (Meinert). Orchesella (Folsom). Muscle C' (distal) corresponds with 9. add. « _ C (proximal) “6 EN Trot. ‘ a D HB «5. pr’t.i. and 6. pr’t. ms. ee ee se 3. ret. rot. and 4. ret., or else 7. rot. and 8. rot. The incompleteness of Meinert’s figure prevents as exact a comparison as is desirable. Japyx is nearest Campodea in structure, and the mandibles of Japyx, which have been described and figured by Meinert (’65), Grassi (’86°), and y. Stummer-Traunfels ('91), are essentially like those of Campodea, but lack the articulated lacinial lobe, there being a lacinial region, how- ever, which (Grassi, 86”, Taf. II., Figura 14) is separated by a trans- verse line from the fulefum. The muscles of Japyx agree with those of Campodea, and it is to be noted that the adductors originate upon a median chitinous plate, or tentorium, just as in Collembola, but not as in Orthoptera. The muscle f of Meinert (’65, Taf. XIV., Figuren 5, 15) has no homologue, it should be said, among the mandibular muscles of Orchesella, and I should be disposed to regard it as an adductor of the head of the first maxilla, had not v. Stummer-Traunfels (91, Taf. I., Figuren 1, 3) figured the tendon of the same muscle in Campodea and Japyx going to the mandible. This author (’91, p. 220) erroneously states that the adductors of Collembola, Campodea, and J apyX are at- tached to the “ Stiitzapparate,” by which he means the lingual stalks (Plate 6, Figure 38, pd.'); these, however, are quite distinct from the tentorium, which he apparently overlooked. Nearly allied to the entognathous genera Campodea and Japyx are the ectognathous genera Lepisma and Machilis. In Lepisma the early 106 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. development of the mandibles, as shown by Heymons (’97?, Taf. XXX., Figuren 13, 15, 17, 20), is simple, and agrees with that of Anurida and Campodea. The finished mandible of Machilis (Oudemauns, ’88, Plate IL., Figuren 25, 26), especially, recalls that of Campodea and Collem- bola by its elongated hollow fulcrum, oblique aperture, basal pivot, distinct head, and (as in Orchesella) well-developed molar surface ; moreover, the adductors originate ona tentorium and are inserted within the mandibles (Oudemans, ’88, Taf. 1, Figur 19; Wood-Mason, 79, p-. 148, Figure 1). Wood-Mason named the apex of the mandible “ ex- opodite” and the molar lobe “ endopodite,” but upon superficial grounds, if one may judge from the evidence of embryology. Both lobes may together represent the endopodite ; but the exopodite, or palpus, is un- represented in the mandible, and it is a secondary lobe of the primary, or stipal, fundament, in the first and second maxilla. Wood-Mason (79) pointed out many interesting similarities which Machilis and Lepisma bear to the most generalized Orthopteran family, the Blattide, and remarked (p. 149), concerning the pivot of Machilis, that ‘the pos- terior ball-shaped condyle of mandibulated insects, clearly foreshadowed in the myriapod, is here fully formed and provided with a distinct neck.” The mandibles of Lepisma, however, more closely approach the Or- thopteran type in being compact (v. Stummer-Traunfels, ’91, Taf. IL, Figuren 5, 6) and partly solidified, and in having broad incisive teeth, a molar surface like that of Orthoptera, and broadly attached adductors. The muscles are said by Oudemans (88, p. 187) to resemble those of Machilis. V. Stummer-Traunfels represents the adductors only, and it may well be that the muscles are really much fewer than in Campodea and Collembola, such a reduction in number, if it occurs, being an approach to the Orthopteran type, in which but two mandibular muscles exist — a stout adductor and a slender abductor. As to the development of the mandibles in Orthoptera, very little has been published. Ayers (’84, p. 241, Plate 18, Figures 20-22) says that in (canthus “the three oral appendages are trilobed ; the lobation is most prominent in the second maxillary and least in the mandibular appendage. The primitive appendage is first divided into two lobes, and the inner of these becomes secondarily divided into two.” The three lobes doubtless represent palpus, galea, and lacinia. Korotneff (85, Taf. XXIX., Figure 6) figures lobed mandibular fundaments for aryllotalpa, In other Orthoptera such lobation has not been recorded. In Blatta, according to Wheeler (’89, p. 348), “There are apparently no traces of lobation in the mandibles.” Packard (’83*, p. 279) says, FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. LOT. “ The mandibles [of Caloptenus] remain single-lobed,” and both Wheeler (93) and Heymons (’95") represent them as simple papilla in all fami- lies of Orthoptera. It may at least be said, however, that the mandibles of Collembola and Thysanura are certainly homologous in their entirety with those of Orthoptera, and hence of all other insects. It is an interesting fact that Heymons (96, Taf. II., Figur 29) dis- tinctly represents mandibular palpi for the larva of Ephemera, — a rare condition ; indeed, Packard (’98, p. 61) terms this appendage of nymphal Ephemerids a “lacinia-like” process, although Heymons states (p. 21) that it is lateral in position, and so figures it. What embryological evidence there is, then, confirms the view based upon anatomical data, that “the mandibles are primarily three-lobed appendages like the maxille ”’ (Packard, ’98, p. 61). Turning now to the Myriopoda, the Symphyla, represented by the single genus Scolopendrella, show marked affinities with Campodea, as is well known. I wish here to emphasize especially the correspondences between the mouth-parts of the two genera, which have never been carefully compared in these respects. Latzel (’84, p. 8, Taf. I., Figur 5) describes the mandibles of Scolo- pendrella as follows : ‘‘ Die Oberkiefer bestehen jederseits aus einer fast horizontal gelagerten, trapezoidalen Chitinplatte, welche am End- oder Kaurande durch eine mittlere Einbuchtung in zwei Partien abgetheilt erscheint, von denen die vordere in vier kraftige, die hintere in vier bis fiinf kleinere Ziihnchen eingeschnitten ist. Eingelenkt sind diese Kie- ferplatten mit dem hinteren und dusseren Eck in eine zwischen Kopf- decke und Unterseite eingelagerte seitliche Lamina, welche einige Aehnlichkeit hat mit der Wange der Insecten und die von Menge als Theil (Stamm) der Oberkiefer aufgefasst wird. Am inneren Hintereck jeder der beiden Oberkieferplatten entspringt eine sehr kraftige Sehne, die in eine betriichtliche Anzahl von Muskelbiindeln auslauft, welche sich unten am Kopfrahmen inseriren.” The mandibles of Scolopendrella therefore resemble those of Cam- podea rather than those of any other insect, in that they are hollow, with a basal (stipal) part articulated to the skull, and a head separated transversely from the fulcrum. The head consists of two primary lobes (galea and lacinia) as in Campodea, but both are movable by muscles, whereas in Campodea the lacinia alone is articulated, and even this no longer has muscular attachments. The tendon and muscles which move the lacinia of Scolopendrella are exactly similar in position and function to the “chitinous rod ” and muscles which adduct the head of the first 108 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. maxilla of Orchesella, Japyx (Meinert, 65, Taf. XIV., Figur 8) and doubtless Campodea. More important, however, is the fact that the tendon of Scolopendrella is comparable with the mandibular retractor (cf. Latzel, ’84, Taf. 1, Figur 5, «, and Meinert, ’65, Taf. XIV., Figuren 5, 15, f, flewor) of Campodea and Japyx, and may be homologous with it. It can be easily understood that, if the terminal lobes in Scolopendrella became immovable by solidification in the mandible, the adductors of those lobes would then serve as retractors of the entire mandible, as in Campodea and Japyx. Grassi (’86*, pp. 15-16, Tav. II., Figure 2, 5) supplements Latzel’s account of Scolopendrella by saying that no true cardo is present, and that the mandible is capable of lateral movements only. Packard (83°, p. 198) says, “The so-called mandibles of the Myrio- pods are the morphological equivalents of those of insects, but structur- ally they are not homologous with them, but rather resemble the lacinia of the hexapodous maxilla.” With the last assertion I do not agree. The mandibles of the more generalized Diplopods are in detail strikingly like those of Scolopendrella (Latzel, ’g4, Taf. 1, Figur 5); for example, those of Polyzonium (Latzel, ’84, Taf. XVI., Figur 203), in which the only fundamental difference is the presence of a cardo in Polyzonium, the stipes, galea, lacinia, and tendon being essentially as in Scolopendrella. The mandible, or protomala (Metschnikoff, *75), of Polyzonium does, in- deed, resemble, not the lacinia, but the entire first maxilla of Thysanura and Collembola. The similarity, however, should not be mistaken for homology ; it rather serves to emphasize the structural agreement of mandibles and maxille,—an agreement which gradually becomes ob- scure in the insect series through the progressive solidification of the mandible, but may nevertheless be traced, as I have shown, from Diplo- podaand Symphyla, through Campodea and Japyx, Machilis and Lepisma, to the more generalized Orthoptera ; thus the differences between the mandibles of Diplopods and Insects are not so great as Packard has affirmed (’98, p. 12). The most that is known about the development of Diplopod mouth- parts we owe to Metschnikoff (’74), who represents only two pairs of oral fundaments, designated “‘ mandibles” and ‘ labium.” -Although this conclusion is also reached by vom Rath (’86), I would not infer with Packard (’83°, p. 199) that there can be only two pairs of oral appen- dages, but would suggest that embryological studies upon the mouth- parts of other Diplopoda may, perhaps, show more. The mandibles, or protomalz, of Chilopoda are generally recognized as FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 109 equivalent to those of Chilognatha, and, indeed, to the mandibles of Hexa- poda and Crustacea. In the mandibles of Scolopendra (Meinert, ’83, Taf. II., Figur 9), for example, there can be recognized cardo and stipes, a distinct head with galeal and lacinial lobes, and even muscles exactly comparable with the adductors and retractors of the mandible in Campodea and Japyx. The affinities of the Chilopods are, however, with the Dip- lopods, — from the stem-form of which they may have developed, — rather than with the Campodeide. Although Packard (98, p. 15) states, “In the Chilopoda also the parts of the head, except the epi- cranium, are not homologous with those of insects, neither are the mouth-parts,” there is really much indirect evidence of homology with the mouth-parts of insects through Diplopoda, Symphyla, and Thysanura, as is indicated above. The mandibles of Crustacea have usually been considered homologous with those of insects. In Malacostraca (Reichenbach, ’86), as in in- sects, the mandibular fundaments are a pair of appendages of the fourth primitive segment. In insects the exopodite (palpus) is absent, but in such generalized groups as Campodea and certain Ephemeride, a “lacinia mobilis” is present ; in Malacostraca the palpus is present, and like- wise, according to Hansen, a similar lacinia is found in the groups Mysida, Cumacea, Isopoda, and Amphipoda, although not in Decapoda. Among insects, the Thysanura most nearly approach Crustacea. Hansen (’93, pp. 205-206) says of Machilis: “Die Mandibeln sind homolog mit denen der Malacostraken ; in Form sind sie denen der Cumaceen 4hnlich, mit einer gut entwickelten, fast cylindrischen Pars molaris, doch ohne Lacinia mobilis; in Einlenkung und Musculatur stimmen sie erstaunend iiberein mit z. B. Diastylis und Nebalia.” Re- ferring to Campodea, Japyx, and Collembola, he remarks (pp. 208-209), “Die Musculatur der Mandibeln ist noch mehr der Crustaceen ihnlich als der Musculatur der JMachilis. Vergleiche Meinert’s Figur von Japyx mit meiner Figur von Diastylis Goodsiri in ‘ Dijmphna-Togtet’ (ich habe nur die drei gréssten Muskeln oder ihre Sehnen wiedergege- ben) oder mit Sars’ Figur von Diastylis sculpta, und man wird betroffen von der erstaunlichen Uebereinstimmung in Form und Richtung der Muskeln und der grossen medianen Muskelplatte.” In conclusion, the mandibles of Apterygota agree in development with those of Orthoptera, but show no trace of lobation except in Campodea, the most primitive form. The mandibles and maxillz are homodyna- mous, and the former are homologous with the mandibles of Scolopen- drella, Crustacea, and probably Diplopoda. 110 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Lingua and Superlingue. Not until Stage 3 are the fundaments of the superlinguze (“ para- gloss” of some authors) observed ; then a ventral aspect of the germ band (Plate 3, Figure 11) reveals two small papille (sw’/ng.) between the mandibles with their centres slightly more anterior than those of the mandibles. Although each small papilla is adjacent or contiguous to the mandibular fundament of the same side, it originates quite inde- pendently ; in other words, it is not the inner branch of a biramous ap- pendage, but a distinct ectodermal evagination, as transections of the germ band (Plate 4, Figure 23, sw’lng.) prove. At Stage 4 (Plate 3, Fig. 12, sw’Ing.) the superlingual fundaments are longer and stouter than before, and have moved back slightly in re- lation to the mandibles until nearly opposite them. At Stage 5 the centres of the superlingue (Plate 3, Figure 21, sw’ Ing.) are behind those of the mandibles, and in cross-sections (Plate 4, Figure 23) the former structures are seen to have exceeded the latter in rate of elongation. The long axes of the superlingue now diverge anteriorly from the median plane and the apices are partly under the mandibles, as in the adult, though the bases retain nearly their original positions in relation to the bases of the mandibles. During this stage is seen the first trace of the lingua (the “ligula,” or “hypopharynx” of some authors), as a slight, median, unpaired, oval, ectodermal evagination (Plate 3, Figure 21, dng.) between the first maxillez. This is the last of the oral fundaments to make its appearance. In Stages 6 and 7 the lingua becomes longer and stouter, and, as seen in a ventral view of the germ band (Plate 5, Figure 30, Ing.), its cross-section is rounded-triangular with its anterior median angle intrud- ing between the two superlingu. Sections show that the lingua and superlinguz have swung forward from their former positions at right angles to the germ band, and that the lingual and superlingual cavities are separately confluent with the general body cavity of the head. In the region of confluence a common cavity —a prolongation of the body cavity —is formed by a median evagination of the germ band itself. In Apterygota the superlinguz, however, never become appendages of the lingua. In ventral aspect, the lingua at Stage 7 (Plate 4, Figure 27 ; Plate 5, Figure 29) is cuneate with rounded apex, and, a little later (Plate 4, Figure 25, Ing.) becomes constricted distally, forming a terminal Jobe. In Stage 8 the lateral surfaces (Plate 5, Figure 34) become concave, FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 111 to correspond with the adjacent convex surfaces of the first maxillee, as in the adult (Plate 7, Figures 44, 45, cht.), and each ventro-lateral edge extends under the neighboring maxilla; in addition, the apex of the lingua becomes separated into two lateral lobes by a median sinus, and the dorsal surface invaginates to form a median longitudinal groove (Plate 7, Figure 42, sul.) ; this lobed condition, however, is quite secondary in origin. The lingua is thickly chitinized, and the hypodermal cells persist in the mature organ. The superlinguz, on the contrary, are but thinly chitinized and at maturity contain no distinct hypodermis cells, except basally, although a complete layer of cells exists in Stage 8. In this stage (8) the superlinguz become triangular in cross-section, as in the adult (Plate 7, Figure 44). Partly on account of the divergence of the superlingue in front, but principally owing to the convergence of the mandibles and maxille, the attenuated distal part of each superlingua becomes situated between the apices of the mandible and the first maxilla of the same side (Figure 44), and the superlingue conform to the adjacent surfaces of the maxille. The most interesting lingual structures are the two basal stalks (Plate 6, Figure 38, pd.’), each of which articulates with the cardo of the same side and also furnishes a firm origin for the adductors and re- tractors of the first maxilla, as in Orchesella (Folsom, ’99, Plate 3, Fig- ure 21). The development of these stalks has never been described. Although difficult to comprehend with a knowledge of the finished con- dition only, it is simpler than might be expected. The key to the understanding of its origin is the fact that each chitinous stalk is formed in a groove which is but a longitudinal evagination of the maxillary pocket, and follows the mesal surface of the first maxilla back to the cardo. The base of the lingual fundament is at da. in Figure 30 (Plate 5), and that of the maxilla at ba.’ ; consequently the stalk is developed in a superficial groove of the germ band itself — that part of the germ band connecting the base of the lingua with the extreme base of the maxilla. In ventral aspect at Stage 7 (Plate 5, Figure 29, pd.’), the continuity of the stalk along the surface of the maxillary pocket is evi- dent. Dorsal to the stalk, of course, the base of the maxilla is connected with the head, but under the connecting region passes the stalk. I must now explain how maxillary muscles become attached to the stalk in spite of the fact that the latter is a superficial formation of the hypodermis. This may be learned from transections at Stages 7 and 8, but also, and more easily, from good serial sections of an adult head, 112 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. such as are shown in Plate 7, Figures 46-50, which successively repre- sent sections in more posterior planes. Figure 46 shows the right maxilla (mz.") sectioned in front of its basal opening and lying free in its pharyngeal pocket ; it also shows the stout, superficial chitinous stalk (pd.') in its hypodermal groove. Figure 47 represents the beginning of an evagination (plz.) of the dorsal wall of the pocket, which grows down between the maxilla and chitinous stalk. Passing back, the intruding hypodermal fold expands, as in Figures 48 and 49 (p/z.), until it almost encloses the stalk. Finally, in the region of the maxillary aperture (Figure 50), and on account of its obliquity, adductor muscles (mw.) are enabled to pass directly from the inner wall of the stipes to the chitinous stalk (pd.’). They are not attached di- rectly to the stalk, but to an intervening cuticula (cta.) ; this, however, amounts to the same thing, because the cuticula and stalk become fused together at about Stage 7, and hardened into a single piece. It should be stated that the hypodermal cells which formed the intervening cuticula, as well as those which formed the stipes, are seen in em- bryological life only; they disappear at the origin and insertion of muscles. At Stage 7 the end of each stalk is already feebly fused with the end of the cardo to form an articulation (compare Plate 4, Figure 25, with Plate 6, Figure 38, ate.). This is a simple process, as both eardo and stipes are superficial and contiguous structures. In the adult Orchesella (Folsom, ’99, Plate 2, Figure 10, /zg.’) a long ligament unites them, and I mentioned a distinct suture as possibly indicating the end-to-end union of two ligaments, which doubtless occurred. The lingual stalks, then, are quite independent of the lingua in origin, except that they are thickened cuticular structures continuous with the lateral cuticula (Plate 7, Figure 45, cht.) of the lingua. When dissect- ing out the lingua at Stage 7, it frequently breaks away from the stalks at the sutures (swt.) shown in Plate 4, Figure 25; these sutures later become obliterated, however. The lingual stalks of Collembola have been mentioned by several authors, for example, de Olfers (62, p. 18) in several genera, Tullberg (72, Taf. IV., Figur 17) in Tomocerus, and v. Stummer-Traunfels (91, Taf. I. Figur 7) in Tetrodontophora. I have seen them myself in all the more common genera; they undergo but little modification within the order. As to the development of the lingua and superlinguz in other insects, very little has been written. Packard (’71, p. 17), as quoted on page ? FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 113 128, did not find the “second maxille” (superlingue) in the embryo of Isotoma. Uzel alone has mentioned the embryonic lingua and super- lingue of Apterygota. In Taf. VI. Fig. 87, he shows, in Tomocerus, three fundaments, which undoubtedly are these structures. In Campodea, happily, Uzel describes with some detail the develop- ment of the “hypopharynx” (’98, p. 35): “Schon in jenem Stadium, bei welchem der Keimstreif sich in seinen mittleren Theilen in das Innere des Dotters einzusenkeu anfingt (Taf. LV. Figur 39), bemerken wir zwischen den beiden Anlagen der Mandibeln zwei einander selr genaherte, ziemlich grosse, flache Platten (hmd.). Diese werden im nachsten Stadium, in dem die Umrollung des Keimstreifs vollendet ist, viel kleiner (Taf. VI. Figur 81, hmd.); dafiir wolben sie sich jedoch bedeutend zu zwei spitzigen Héckern vor. Bald erscheint zwischen den Anlagen der ersten Maxillen eine unpaare, grosse, flache Platte (Figur 82, hmx.,), vor der man eine kleinere sieht. Letztere befindet sich zwi- schen den beiden vorher beschriebenen spitzigen Héckern und gehort noch dem Mandibularsegmente an (Figure 83, hmd.'). Die unpaare, dem ersten Maxillarsegmente angehdrende Platte schickt sich nun an, iiber die beiden Hocker und die zwischen denselben gelegene kleine Platte vorzuwachsen (Figur 84), und zwar etwa in der Zeit, zu welcher das Thier ausschliipft.” After hatching, continues Uzel (98, p. 48), “ Von den drei schon friiher beschriebenen Héckern, die zwischen den beiden Anlagen der Mandibeln lagen, wird der mittlere immer kleiner. Bei dem erwachsenden Thiere haben sich die beiden seitlichen zu runden bewimperten Schuppen ungebildet, welche von Meinert (’67) als Para- glossze bezeichnet worden sind. Zwischen denselben befindet sich der nun sehr klein gewordene mittlere Hocker als unbedeutendes Gebilde, welches die beiden seitlichen Schuppen verbindet. Die gréssere, zwi- _ schen den beiden Anlagen der ersten Maxillen gelegene Platte hat sich auch in eine, aber entsprechend der michtigeren Anlage, grdssere Schuppe verwandelt und ist iiber die beiden Schuppen des Mandibular- segmentes erst beim geschlechtsreifen Thiere giinzlich vorgewachsen (Taf. VI. Figure 85, hmz.,). Sie stellt Meinerts Ligula vor. Sowohl die von Meinert (’67) als Paraglosse, als auch die von demselben als Ligula gedeuteten Theile sind, wie wir gesehen haben,*ihrer Anlage nach als Hypopharynx aufzufassen.” The “hypopharynx ” of Campodea is, then, undoubtedly homologous with the lingua and superlingue of Anurida, with the development of which it fundamentally agrees. In Anurida, however, as contrasted with Campodea, the superlingual fundaments do not show the early 114 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. decrease in size, and a small median lobe does not appear on the anterior surface of the lingua. In the finished condition in Campodea (Meinert, ’65, Taf. XIV. Fig- uren 17, 19) lingua and superlingue are simple but distinct lobes, and the small fourth lobe mentioned by Uzel persists. The lingual stalks are surprisingly like those of Orchesella ; the articulation with the cardo Meinert did not show, but it has since been observed by v. Stummer- Traunfels. The English translator of Meinert’s paper is really responsible for the use of the terms “lingua” and ‘paraglosse”’ in connection with this subject, and not Meinert himself; the latter writer applied only the Danish expressions ‘“‘ Tungen ” and “ Bitungens tvende Flige.” Von Stummer-Traunfels (91, Taf. I. Figur 11) also represents the “ Ligula,”’ “ Paraglosse,” and “Stiitzstiicken ” of Campodea. On page 121 I criticise this author for holding that the so-called maxillary palpus of Collembola belongs to the neighboring superlingua. The em- bryology shows that the delicate membrane connecting either palpus and superlingua is of quite subsidiary importance, being simply as much of the cuticula of the maxillary pocket as intervenes between the base of a superlingua and the adjacent maxilla, —in fact, only the anterior portion of the cuticula surrounding the tissues which attach the maxilla to the head. Japyx agrees closely with Campodea in the structure of these organs (Meinert, ’65, Taf. XIV. Figur 8; von Stummer-Traunfels, ’91, Taf. I. Figur 10), and there is no doubt about the homology of the lingua, superlinguee, and lingual stalks of Japyx with those of Collembola. In the words of v. Stummer-Traunfels (’91, p. 221), ‘ Diese typische Form des Stiitzapparates und der Befestigung der Cardines an diesem findet sich bei Campodea, Japyx und den Collembola in beinahe identischer Weise ausgebildet.” The author is mistaken (91, p. 222), however, in saying that the mandibles are attached to the Stiitzapparate, apparently having overlooked the tentorium, which is quite another structure than his “ Stiitzapparat.” Regarding Lepisma, Heymons (’97', p. 595) simply remarks: “Ich . . . bemerke nur, dass die Bildung der einzelnen Korpertheile, z. B. des Hypopharynx der Mundwerkzeuge, durchaus an den bei Orthopteren bekannten Typus anschliesst.” Machilis, also, has decided Orthopteran affinities, as Wood-Mason (’79) found, yet the mouth-parts of both Lepisma and Machilis, although ectognathous, as in Orthoptera, are constructed upon fundamentally the FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 115 same plan as those of the entognathous Apterygota. The similarity is evident in part from the following account of Machilis by Oudemans (88, p. 186): “ Letztere [Ligula], Figur 28 Zz, reicht mit ihrem freien Ende ungefiihr ebensoweit als die Unterlippe und wird durch zwei Chitinstiibchen gestiitzt, Figur 28 S, Figur 3058. Mit der Ligula sind noch zwei Sticke, Figur 30 P, verbunden, die ich als Paraglossee auffassen mochte. Sie sitzen an einer Chitinleiste, die sich auf der Dorsalseite der Ligula findet. Jede Paraglossa ist an ihrem freien Ende noch einiger- massen vertheilt (ich glaube in drei Lobi) und hat einen kleinen Vor- sprung an ihrer Basis, Figur 30 A. Es scheint mir, dass die Paraglossz ausserdem noch festsitzen an den Stiitzstiickchen der Ligula, Figur 3058.” . . . “Die Maxillarspitzen treffen einander mithin in dem Zwischen- raum zwischen Ligula und Paraglosse, Figur 21, die Mandibularspitzen zwischen Paraglosse und Labrum.” Von Stummer-Traunfels (’91) repeats some of Oudemans’ figures of Machilis. In Machilis, I find that the first maxille articulate with the skull— no longer with the lingual stalks —and the stalks, although evident, are much reduced and apparently functionless. The salivary glands open, as in Orthoptera, under the base of the lingua. In Orthoptera, the most generalized of the Pterygota, there is a well- developed hypopharynx, or lingua, which exactly corresponds in position with the lingua of Apterygota, being a median papilla between the bases of the first and second maxille. In Periplaneta (Miall and Denny, ’86, p- 127, Figure 71) it is borne upon two chitinous stalks, clearly com- parable with those in Apterygota. Looking for traces of superlingue in Melanoplus femoratus, I found them, as large dorso-lateral rounded lobes, intimately united, however, with the lingua. This union is already foreshadowed in Machilis and Lepisma. I also found — almost accidentally — two rudimentary, chitinous, divergent stalks, extending back into the head from the ventro-lateral regions of the base of the lingua. The significance of these facts is clear, although the meaning of the lingual appendages, which have apparently been overlooked or disregarded in most Orthoptera, could hardly have been ascertained without studying the less specialized Apterygota. In Packard’s figure of Anabrus (798, p. 73, Figure 71), also, the lingua and left superlingua are evident. In the rare and singular Hemimerus, Hansen (94, pp. 70-71, Plate 2, Figures 9, 10, 4.) finds a “ hypopharynx ”’ and “ maxillule,” as well as 116 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. chitinous stalks, all of which distinctly are as in Collembola, Campodea, and Japyx, except that the superlingue of Hemimerus appear to be fused with the lingua. Figure 10 of Hansen bears a close resemblance to my Figure 27 of Anurida, although Hansen says (p. 87), “ especially the structure of the mouth removes it [Hemimerus] very far [?] from the Thysanura and leads it to the Orthoptera.” In the young larva of Ephemera, Heymons (’96, p. 22, Taf. II. Figur 29) finds that ‘Der Hypopharynx entsteht . . . auf ahnliche Weise wie bei den Orthopteren. Auch an ihm findet eine Art Gliede- rung statt, dergestalt, dass von der eigentlichen Hauptmasse zwei laterale vordere Zapfen abgetrennt- werden, die mit kleinen Hiarchen bedeckt sind, wihrend der eigentliche Hypopharynx am Ende einen Besatz von feinen (Sinnes-) Borsten tragt.” His figure of lingua and superlingue might fairly represent those structures of Anurida in Stage 7 (Plate 4, Figure 27). In the imago the mouth-parts are, of course, atrophied. In another Ephemerid nymph, Heptagenia, Vayssiére (’g2, pp. 113- 114, Planche 5, Figure 46) found a highly developed lingua, or hypo- pharynx, fused with large lateral pieces [superlinguee] and suggests that they indicate a distinct primitive segment, —a possibility which will be discussed later. He states (p. 106), ‘‘ La langue ou hypopharynx.. . est assez dévelopé chez tous les individus de la famille des Ephémérines, & exception du Prosopistoma, ow il est tres rudimentaire.” I shall not cite descriptions of the “hypopharynx”’ of additional in- sects, because I have nothing more to add, and the subject has been well treated of by Kolbe (’$0, pp. 213-217, Figuren 126-134), Packard (98, pp. 70-83, Figures 70-87), and others. Packard’s comparative — account, in particular, is most excellent and well illustrated. (In his Figure 69, by the way, the abbreviations p. and hyp. should be inter- changed.) Briefly, the lingua is found in every order of insects, and although highly specialized in suctorial orders, retains, nevertheless, the — same position and nearly the same relations to the salivary duets that it does in the more generalized mandibulate orders which I have de- scribed. It is an interesting fact that in the Lepidopterous genus, Micropteryx, Walter (’85, Taf. XXIV. Figur 11) shows two hypo-— pharyngeal stalks, readily comparable with those of Apterygota. The superlinguze — which, as I have shown, originate quite indepen- dently of the lingua in Apterygota, but become more or less united with it in Orthoptera and Ephemerida— should hereafter be recognized as morphologically important structures, and be searched for in even the most specialized haustellate orders as more or less intimate constituents FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. Lay of the *‘ hypopharynx,” which term, then, may refer collectively to the lingua and “superlingue.” The necessity for this new term, also brought out on page 132, will appear from the following synonymical table : — AUTHOR. APTERYGOTA. HYPOPHARYNX. Lingua. Superlingue. De Olfers, 62 Collembola lingua organa cochleariformia Meinert, ’65 Thysanura tungen bitungens tvende Flige se (trans.), °67 Thysanura lingua paraglosse Packard, ’71 Collembola <<, os second maxillz Tullberg, ’72 Collembola lamina hypopharyngis laminz hypopharyngis ; inferior superiores Lubbock, ’73 Collembola and P Thysanura ligula, lingua second maxillz Grassi, ’86 Thysanura ligula paraglossze Oademans, ’88 Thysanura ligula paraglossze VY. Stummer-Traunfels, ’91 Collembola and Thysanura ligula paragloss~ Hansen, ’93 Collembola and Thysanura hypopharynx maxillulze Heymons, ’97 Lepisma hypopharynx Uzel, ’98 Collembola and Thysanura hypopharynx Folsom, '99 Collembola glossa paraglossz Among Pterygota, the term “ hypopharynx” of Savigny is fixed in application, although the compound nature of the organ is not gener- ally known. Synonymous with “ hypopharynx ” are the following terms (see also Packard, ’98, p. 71): lingua (Savigny, ’16), ligula (Kirby and Spence, ’28), langue ou languette (Dugés, ’32), lingua (Westwood, 739, p. 9), tongue (Taschenberg, 79), hypopharynx (Dimmock, ’81; Burgess, ’80, and most others). “Ligula,” “glossa,” and “paraglosse” are terms established in Pterygota, but less fixed in the little-known Apterygota, and therefore more easily discarded in the latter group, as advised on pp. 132-133. “ Maxillule ” and “second maxille” as applied to superlingue are un- fortunate because based upon unproved homological assumptions. The need for a new term, then, becomes evident. I have therefore suggested “ superlinguc.” In Scolopendrella authors have omitted to mention whether the hypo- pharynx is present or not. Referring to Diplopoda, however, to which Scolopendrella is most nearly related, Packard (’98, p. 13) says, “The hypopharynx, our ‘ labi- ella’ (Figure 6), with the supporting rods, or s¢éli linguales (sti. 1.), of Meinert, are of nearly the same shape as in some insects.” Latzel (Taf. IX. Figur 104; Taf. VI. Figur 72) represents “ein Zwischenstiick der VOL. XXXVI.— NO. 5 3 118 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Zunge,” for Lysiopetalum and Craspedosoma respectively, as well as two lateral lobes, or “ Zungenlappen ” (lobi linguales). These structures, although united with the gnathochilarium, are probably homologous with the separated lingua and superlingue of Apterygota, but, in the absence of the necessary embryological investigations, that is all that may be said. In the Chilopoda no structure analogous to the hypopharynx appears to be known. The “superlingue ” of insects are homologous with the first maxille of Crustacea. In Anurida I have found (Plate 4, Figure 28, su’lng.) a distinct primitive ganglion — the fifth — for the superlingue, represent- ing the fifth, or first maxillary, ganglion of decapod Crustacea. This ganglion is eventually incorporated with the subcesophageal ganglion, and no superlingual nerves develop. Moreover, the superlingue origi- nate between the mandibles and so-called “ first maxillee” of Anurida. The superlingual fundaments, however, never become biramous — an exopodite or palpus does not appear —and are not segmented, like the Crustacean first maxilla. In fact, they are much reduced structurally and functionally in Apterygota, and gradually reduced to disappearance in ascending the Pterygote scale. Hansen (’93) regarded the superlinguze — or “ maxillule,” as he termed them — from their position, as equivalent to the Crustacean first maxille, emphasizing the opinion of v. Stummer-Traunfels (91) that the super- linguz bore palpi. The latter argument cannot be used, however, be- cause, as I show (p. 121), the palpi in question belong to the “ first maxille.” The lingua, usually termed “hypopharynx ”’ among insects, may easily be homologized with the hypopharynx of Malacostraca. It originates quite independently of the superlingu as a median, unpaired papilla, is not supplied with a primitive ganglion or distinct nerves, and can, no more be regarded as a distinct segment than can the labrum. In Orchesella and Anurida it finally becomes distinctly bilobed by a median groove, but the bilateral condition is clearly secondary. Packard’s evi- dence (98, pp. 82-83) that the hypopharynx is “composed of, or sup- ported by, two bilaterally symmetrical styles both in Myriapods and in insects” has little weight, in view of what I have found to be the devel- opment of these “ lingual stalks,” The hypopharynx of insects, then, is a compound structure, the com- ponents of which originate independently. The median ventral lingua, like the labrum, does not represent a pair of appendages; the dorso- FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 119 lateral “superlingue,” which have been usually overlooked or disre- garded in Pterygote insects, represent a distinct though reduced somite, as confirmed by the presence of a primitive ganglion. The superlinguz are homologous with the first maxilla of Malacostraca, and are probably represented in Diplopoda. The lingua of insects is homologous with the Crustacean hypopharynx and probably with the median constituent of the gnathochilarium of Diplopoda. Maxille. The fundaments of the “ first maxille ” appear next after those of the mandibles, and at Stage 1 (Plate 1, Figure 1 ; Plate 2, Figures 8, 8a, mx.") are a pair of small hemispherical papillz, similar to those of the mandibles. At Stage 3 they are longer than the mandibles and must consequently have lengthened faster. As seen in transections of the germ band, the maxilla is at first a simple ectodermal evagination, api- cally rounded, but at Stage 3 (Plate 3, Figure 15) the apex is flattened, and a lateral lobe, the beginning of the palp, has appeared ; this lobe is also seen in the ventral aspect of the germ band (Plate 3, Figure 11, pip.) as well as in the lateral views (Plate 2, Figures 9, 10). The pos- _ terior aspect of the left first maxilla when dissected out is given in Plate 3, Figure 17. At Stage 4 (Plate 3, Figures 12, 19) the maxilla has elongated con- siderably and its base is covered by the lateral fold of the germ band (Plate 3, Figure 19, pli. or.), as already mentioned. In the following stage (Plate 3, Figures 20, 21, mz.') the maxilla and palpus, though longer, are more nearly concealed by the lateral fold. The form of the maxilla with its palpus at this stage is shown in Figure 22, which was drawn from a dissection; the base of the maxillary fundament is already oblique, precisely as described for the mandibles, and the first maxille have begun to converge toward the median plane. It is to be remem- bered that the palpus is here a secondary lobe of the primary fundament. At Stage 7 (Plate 4, Figure 24; Plate 5, Figures 29, 30, mx.1) the first maxillz, now covered by the lateral folds, have swung forward through an angle of almost ninety degrees (Figure 24), like the man- dibles. Claypole (98, p. 263) states that “a flexure of the embryo be- gins that results in crowding the mouth-parts together to form a definite head,” but such a purely mechanical interpretation will not serve, be- 120 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. cause the paired mouth-parts are still at right angles to the germ band long after involution has occurred (Stage 5, Figures 5, 20). During © this stage (7) the first maxille are attenuated toward their free ends; a ventral view (Plate 5, Figure 29, mx.) shows their position in rela- tion to the lingua, and the extent of their convergence toward each other. The maxilla is quite unattached to the pharyngeal pocket (Plate 7, Fig- ures 44-50, cav. buc.) in which it lies, except where the margin of its basal aperture becomes confluent with the wall of the pocket (Figure 50) ; it has the form of a modified cone with an oblique, dorso-mesal, basal — opening, as shown in transections (Figures 44-50, ma.!). The parts — named stipes and chitinous rod in my paper upon Orchesella are, as I have since found in that genus and in Anurida, distinguished simply by a greater deposition of chitin, and are connected above and below by delicate chitinized membranes, which I did not recognize until influ- enced by embryology to search for them. The “ chitinous rod,” then, is proved both by anatomy and development to be but a part of the stipes. During this stage (7) certain important differentiations of the first maxillz are observable, if those organs are dissected out. The articula- | tion between stipes and cardo (Plate 6, Figure 38, atc.) appears super- — ficially as a notch, and in frontal section as a’ less chitinized region, as might be expected ; in a small hypodermal pocket is formed the stipal projection (Figure 38, p7j.) noticeable in the finished organ. The cardo, now transverse in position, was formerly the basal region of the lateral surface of the primary maxillary fundament, before the basal attachment became oblique. The articulation between the cardo and lingual stalk was described on page 112. In this stage, too, the head of the maxilla becomes vaguely separated from the stipes by a constriction (Plate 5, Figure 29). Later, the con- striction is more pronounced (Plate 4, Figure 25), and the apex of the head is fashioned into an acute curving lobe, — the fundament of the galea (Plate 4, Figure 26, ga.) or “aussere Lade.” The “head” is lined with a continuous layer of hypodermis cells. Next, on the mesal side of the head, a second lobe appears, the lacinia (/en.), or “ innere Lade.” Both galea and lacinia, then, become toothed on the mesal face, the teeth of the latter being produced each by a single cell ; the larger teeth of the former are secreted each by many cells. Eventually (Plate 6, Figure 39) the galea (ga.) becomes thickly chitinized except for a central hollow core, but the lacinia (/en.) remains thinly chitinized even in the adult. As in the mandibles, the hypodermis is finally excluded FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 121 from the head of the maxilla, but through an opening in the constricted region nerve fibres may be traced to the lacinia. At this stage (7) the first maxillary palpus (Plate 5, Figure 30, pip.), though still present, is no larger than it was in Stage 5 (Figure 22, plp.). In the newly hatched insect no trace of this palpus exists, hence it must have been resorbed. In the adult Orchesella, on the contrary, the palpus is functional and highly developed ; other facts also indicate that Anurida is a degraded form. Von Stummer-Traunfels (’91, p. 226, Taf. I. Figuren 6, 7), following Tullberg (’72, Taf. IV. Figur 17), observed a connection between the maxillary palpi and the so-called paraglosse of Collembola, and makes use of this union (p. 226) as the first of his reasons for recommending an im- proved designation of the mouth-parts, in the following words: “I. Die grosse Unwahrscheinlichkeit, dass der sogenannte Maxillartaster der Col- lembolen wirklich zur Maxille gehGrt, indem diese von jenem vollstandig getrennt ist und derselbe vielmehr in innigem Verbande mit den Para- glossen steht.” Hansen (’93, p. 209) uses this conclusion in proving that the “paraglosse” of Collembola and Thysanura are homologous with the first maxille of Crustacea. Without discrediting his conclu- sion, I have already shown (Folsom, ’99) upon anatomical data the trivial nature of the union between palpus and “ paraglossz ” (super- lingue). I have now proved upon embryological evidence (Plate 3, Figure 22) that the palpus belongs to the maxilla, and have also shown (p. 114) that the chitinous membrane connecting it with the superlingua ‘is simply incidental, and is only that part of the wall of the maxillary pocket which necessarily intervenes between the first maxilla and super- lingua of the same side. The fundament in Isotoma designated first maxilla by Packard (’71, Plate 3, Figure 13) is undoubtedly, from its position in relation to the first pair of legs, second maxilla; therefore what he regards in the same figure as a mandible must be a first maxilla. Ryder (’86) followed Packard in this matter, but Wheeler (’93, p. 57, Figure 6) shows the fundaments in their proper position. Claypole (’98) correctly identifies the first maxillary fundaments in Figures 43, 46, and 47, but does not mention the palpus. Uzel (’98) gives a figure of the first maxillary fundaments of Tomo- cerus and remarks (p. 36): ‘In jenem Stadium, bei welchem die Um- rollung des Keimstreifs vollendet ist, bemerken wir, dass sich die Anlagen des ersten Maxillenpaares (Taf. VI. Figur 87, mz.,) in zwei Hocker getheilt haben, und zwar in einen fusseren linglichen und in einen inneren 122 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. stumpf dreieckigen. Der innere Hocker diirfte nach Analogie mit Campodea den lobus internus, der aiissere die gemeinschaftliche Anlage des Lobus externus und des Palpus maxillaris vorstellen.” I entirely disagree with the author as to the interpretation of the lobes. In Anurida the lobus externus is not developed out of the palpal lobe of the biramous fundament, but the remaining lobe is the common funda- ment of lobus externus and lobus internus. Therefore Uzel’s foot-note on page 36, “An den Maxillarpalpen von Macrotoma (Tomocerus) vul- garis fand ich selbst einen kleinen Vorsprung, der wohl als Lobus ex- ternus zu deuten ist,” etc., is open to criticism; the minute papilla to which he refers is precisely like several other papillee distributed upon the palpus (see Folsom, ’99, Plate 3, Figure 18, plp.), except for a trifling difference in size. It is very doubtful if a difference in this matter exists between Anurida and Tomocerus, especially since the process as observed by me agrees with that of insects in general, as far as is known, excepting possibly Lepisma, presently to be noticed. . Uzel applied to Tomocerus conclusions drawn from Campodea, in which he (’98, pp. 33-34, Taf. VI. Figuren 79, 80) derives the galea from the palpal lobe. His diagrams, unfortunately, do not elucidate the basal relations of the three principal lobes: palpus, galea, and lacinia. The completed first maxillee of Campodea (Meinert, ’65, Taf. XIV. Figuren 17, 18; Grassi, ’86°, Tav. IV. Figure 2, 13; v. Stummer- Traunfels, ’91, Taf. I. Figuren 5, 11) are remarkably like those of Col- lembola (Folsom, ’99, Plate 3, Figures 18-21): the cardo is articulated to the superlingual stalk in the same way; the hollow stipes, distinct head, galea, and lacinia are also alike, and resemble less the homologous parts of Pterygote insects. The solid bifid galea and the fringed seven- lobed lacinia of Campodea, as I call them, are by Grassi and vy. Stummer- Traunfels regarded collectively as the ‘‘innere Lade” or lacinia. The latter author says (p. 223), ‘“ Man kann daher bei den drei vorliegenden Formen eine successive Riickbildung des Aussenladens annehmen. Bei Japyx noch zweifach gegliedert, ist er bei Campodea schon mebr reducirt und fehlt bei den Collembolen giinzlich.” It is curious to observe how authors have followed one another in deriving the galea from the palpal fundament. I have shown in Anurida (anticipating later conclusions) that the galea and lacinia both originate from the “ endopodite” of the bifid fundament. Japyx, of course, agrees substantially with Campodea. The second maxilla, however (Meinert, ’65, Taf. XIV. Figuren 8, 9; Grassi, ’g6°, FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA, 12a Tay. II. Figure 2, 3, 6, 8; v. Stummer-Traunfels, ’91, Taf. I. Figuren 4, 10) has a two-lobed galea and a four-lobed lacinia. In Japyx, thanks to Meinert’s figure (65, Taf. XIV. Figur 8), the muscles may be clearly homologized with those I (99, Plate 3, Figures 20, 21) have described for Orchesella. As Meinert did not designate the muscles, I can simply say that they severally correspond with those labelled by me 3. add., 4. add., 10. add., 7. add. or 9. prt. add., and one muscle with both 3. pr’t. add. and 6. pr’t. add., while one of two others probably represents 8. ret. add. In Lepisma, according to Heymons (’97*, p. 592, Taf. XXX. Figuren 13, 13, 17, 20), the fundament of the first maxilla forms the pal- pus, at the base of which appears a mesal lobe, which itself divides to form galea and lacinia. This account is, then, at variance with mine on Anurida, that of Uzel on Campodea, and that of Ayers for the Orthop- teran genus (Ecanthus, and is, so far as I know, unsupported by the re- sults of other authors. In fact, Figure 13 of Heymons even suggests that the palpus is a lateral lobe of the primary fundament, as I have found it to be in Anurida. As to the origin of the three first maxillary lobes, Uzel, Heymons, and myself disagree, as I have said. Uzel’s account agrees with mine, in so far as he makes the palpus a lateral evagination of the primary, or stipal, fundament ; and Heymons, like myself, derives both lacinia and galea from the inner lobe of a biramous appendage. In its final form, the first maxilla of Lepisma is easily recognized as homologous with that of other Thysanura, but approaches remarkably the same organ in Orthoptera, especially that of the Blattide. As in other Apterygota, the stipes (v. Stummer-Traunfels, ’91, Taf. II. Figur 11 ; Muhr, ’77, Taf. VII. Figur 45) has a basal opening, cardo, distinct head, galea, and lacinia, and the origin of the muscles (Oudemans, 88, p. 187) “findet man auch hier an Chitinstiicken im Kopfe.” The palpus in Lepisma, however, is five-jointed, as in Orthoptera. What I call galea and lacinia are also, in this particular case, named ‘“ Aussen- lade” and “ Innenlade” by v. Stummer-Traunfels. Machilis is nearer than Lepisma to Campodea and Collembola in the structure of the first maxille. As may be seen from the figures by Oudemans (’88, Taf. II. Figur 27) and v. Stummer-Traunfels (91, Taf. II. Figuren 8, 9, 10), the positions of the cardo, stipes, galea, lacinia, and palpus are exactly comparable in the three groups. The palpi in Machilis, to be sure, are seven-jointed, and a palpiger is present, asin Orthoptera. The structure identified by v. Stummer-Traunfels as 124 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. ‘“Aussenlade” in Machilis, cannot be homologous with the part bearing the same name in other insects, for in Machilis it is clearly a part of the palpiger instead of being a constituent of the head of the maxilla. The two adductor muscles described by Oudemans (’88, Taf. I. Figur 19) as extending from the inner wall of the maxilla to a median tentorium, are probably the homologues of 4. and 6. pr’t. add. of Orchesella (Fol- som, 799, Plate 3, Figure 20). In Gicanthus, Ayers (84, p. 241, Plate 18, Figures 20-22; Plate 19, Figure 5) has traced the development of the first and second maxille as far as the trilobed condition, his ideas (p. 241) agreeing with mine on Anurida: “ The primitive appendage is first divided into two lobes, and the inner of these becomes secondarily divided into two.” Patten (’84, p- 596) says, “A rather striking variation was found in the first and second maxilla of Blatta, which were formed respectively of two and three lobes.” Wheeler (’89, p. 348) adds, regarding the same genus, | “The outer of the three lobes of each maxilla becomes the palp, while the inner two become the galea and lacinia of the adult.” Heymons (95°, p. 19) states that in Forficula, ‘drei selbstindige Aeste zu erkennen sind, aus denen Lobus internus (lacinia), Lobus externus (galea) sowie . der Palpus hervorgehen.” This trilobed stage is exactly comparable with that of Lepisma, although Heymons and Ayers differ as to its derivation. Although Wood-Mason (’79) emphasizes the agreement between Machilis and Orthoptera, I may say that Lepisma is intermediate between the two in structure, with decidedly orthopteran affinities. Especially is this true of the first maxille. The cardo, stipes, galea, lacinia, and palpus of Lepisma (Muhr, ’77, Taf. VII. Figur 45, or v. Stummer-Traunfels, ’91, Taf. II. Figur 11) not only agree in position with those of Blatta (Muhr, ’77, Taf. II. Figur 12, or Packard, ’83*, Plate XXVIII. Figure 12, Periplaneta), but exhibit a surprising agree- ment in form, as well as the number of palpal segments ; in both groups, also, a palpifer is differentiated. Through Lepisma, therefore, the first maxille of Collembola may be homologized with those of Orthoptera, and hence all other Pterygote orders. I have traced the homologies, part for part, between Lepisma and all the families of Orthoptera, as well as the genera Ephemera, Myrmeleon, and Corydalus, in which lat- ter genera the nymphal first maxille are but little specialized in form, Heymons (’96, p. 19) states that in the Libellulid genus Epitheca, “ Erst spiiter gliedert sich von der Aussenseite der Maxille eine kleine rundliche Erhebung ab, welche die Anlage des Tasters darstellt (Figur FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA, 125 19, palp. mx.'), wibrend das in der directen Fortsetzung des urspriing- lichen Maxillen-Zapfens liegende Endstiick zur Lade (lobus) wird.” This agrees with Anurida, Campodea, and (canthus, but disagrees with the account given by Heymons himself for Lepisma. Turning to the Myriopods, Scolopendrella, while undoubtedly more closely allied to the Diplopods, nevertheless shows in many ways inter- esting correspondences with Campodea, as other writers have already stated. The lateral parts of the plate termed the “ gnathochilarium ” resemble in several respects the first maxille of Campodea. According to Latzel (84, Taf. I. Figuren 6, 7) and Grassi (’86*, Tav. II. Figure 5, 10), there is an elongated hollow stipes bearing an outer (galeal) and also an inner (lacinial) terminal lobe, both of which agree in detail with the comparable structures of Campodea and Japyx; for in Cam- podea, a one-jointed palpus is present, and in Campodea, Japyx, and “Collembola, a “chitinous rod” extends backward from the lacinia. The few muscles shown by Latzel (’84, Taf. I. Figur 7) are to be compared with 5. and 6. pr’t. add., and 8. ret. add. of Orchesella (Fol- som, 99, Plate 3, Figures 20, 21). Grassi (’86*, p. 16) states that muscles from within the organ pass to an endoskeleton, which, as one may see from his Figure 25, is essentially like the “ lingual stalks” that I have found in Orchesella and Anurida, and still more nearly like the same structure of Campodea and Japyx. All these similarities confirm the view, based primarily upon other anatomical data, that Scolopendrella most clearly represents the hypothetical ancestor of insects. Among Diplopods the passage from the more generalized genera, as Lysiopetalum or Craspedosoma, to Scolopendrella is clear. In the first genus, especially, are seen a cardo (not described as yet for Scolopen- drella), stipes, galea, and lacinia, all simple in structure, but no palpus. I should state, however, that it remains to confirm these homologies by embryology. In Campodea the second pair of jaws is usually homologized with the first maxille of Insects; but, except in position, there is little re- semblance between the two organs. The first maxillze of insects are usually homologized with the first maxillz of Crustacea, but if, as I maintain, the ‘“ superlingue” are equivalent to the latter organs, it follows that the hexapod first maxille correspond to the Crustacean second maxille. The primitive biramous character of Crustacean mouth-parts is well known, and Hansen (’93, p. 198) has, in connection with this subject, 126 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. formulated a significant law — “dass man drei Glieder im Stamm yon allen gespalteten Gliedmassen bei den Crustaceen als ein primiires Verhiiltniss annehmen muss, und diese Zahl hat sich, wenigstens in den angefithrten Fillen, deutlich erhalten.” In fact, Hansen himself (p. 206) has homologized the first maxillee of Machilis with the second maxille of Crustacea, on account of the three axial segments and the position of the palpus, saying: ‘Der Bau der Maxillen ... stimmt also genau mit den Maxillen der Eumalacostraken.” The axial segments of the Crustacean appendage are on this view successively equivalent to cardo, stipes, and palpifer of Hexapoda. It must be admitted that these anatomical agreements, if appealed to alone, may logically be used to support other views than my own, since all the Crustacean appendages are constructed upon the same plan; but the equivalence of the neuromeres in Hexapoda and Crustacea is a mat- ter of the greatest significance. Viallanes has proved that the first three neuromeres in the two groups agree in great detail, and I find that his con- clusions apply equally well to the succeeding neuromeres, It is very sig- nificant that in most cases the appendages of equivalent somites have the same function in the two groups, and that all the paired nerves of the head in Collembola agree exactly in position with those of decapod Crustacea. Summarizing: The first maxillee of Apterygota develop in all essential respects like those of Orthoptera, with which they may be homologized in detail. In Anuridaa palpus appears, but is resorbed before hatching, indicating the descent of Anurida from a form in which the first maxil- lary palpi were functional. The first maxille of Campodea are clearly to be homologized with those of Scolopendrella, and less clearly with the lateral portions of the Diplopod gnathochilarium. The first maxilla of Hexapoda pass through a biramous stage, such as obtains among Crustacea, are comparable with Crustacean second maxille in some detail, and are homologous with those of Malacostraca. Labium. The fundaments of the labium, or “second maxille,” appear next after those of the first maxille, and at Stage 1 (Plate 1, Figure 1; Plate 2, Figure 8, mz.?) are a pair of simple conical elevations rising perpendicularly from the germ band and slightly longer than the funda- ments of the mandibles and first maxille. In the following stage (2) they are longer and more cylindrical (Figure 2); in Stage 3 (Figure 3) FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 1i3F they are somewhat larger, and ventral or lateral surface views of the germ band (Plate 2, Figures 9,10 ; Plate 3, Figure 11) disclose a distinct lateral lobe, the palpus (p/p.), which is larger than that of the first maxilla. A second maxilla, as dissected out at this stage, is shown in Figure 18 (Plate 3). Transections of the germ band (Plate 3, Figure 14) show the palpus to be an outfolding of the antero-lateral face of the primary maxillary evagination, just as in the case of the first maxilla. At this stage (Plate 2, Figures 9, 10) there appears near the mandibles a lateral evagination (plz. or.) of the germ band, destined to form the side of the face; this fold grows backward until it involves the base of the second maxilla of the same side, and the internal cavities of the two folds become one. In Stage 4 (Figure 4; Plate 3, Figure 19) it has already involved the base of the second maxilla; the apex of the maxilla, however, is still free from the fold, and the palpus (Plate 3, Figure 12, pip.) is as large as that of the first maxilla. At Stage 5 the second maxille (Figure 5; Plate 3, Figure 20, mz.) are long, oval in cross-section, and project at right angles to the germ band; the antero-lateral region of the base is confluent with the mouth- fold (Figure 21). At this period all trace of the second maxillary palpus becomes Jost; it has not become involved in the mouth-fold, which is still restricted to the base of the maxilla, but has been rapidly resorbed and appears at last as indicated in Figure 20. In the next stage (6) the second maxillz (Figure 6) converge toward the median plane like the other pairs of oral organs, and similarly swing forward. At Stage 7 (Plate 2, Figure 7) the second maxille and mouth-folds are quite confluent (Plate 5, Figures 30, 34), but the anterior part of each maxilla is still distinguishable as a swollen lobe, or less flattened region (Plate 4, Figure 24, /ab.). The bases of the second maxilla, although widely separated in Stage 5 (Plate 3, Figure 21, mz.*), sub- sequently spread toward the median plane, become thinner, and gradually form a single plate; the median sinus between them shortens until the condition shown in Plate 5, Figure 29 (/ab.) is attained. The union of the second maxille with each other is not a simple contact and fusion resulting in a median suture ; but a confluence of the cavities of the two maxillz occurs and progresses forward (7. ¢., distally), ceasing, however, before obliterating the median sinus, which remains in the adult (Plate 7, Figures 43, 45, sut. m.and sul.). Although the finished labium bears a median ventral groove, the groove does not indicate the fusion of the fundaments; at Stage 7, when the labial plate is complete, no trace exists of the groove, which is formed in a later stage. A comparison of 128 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. my figures (Plate 3, Figure 20; Plate 5, Figure 29) will show that practically the entire ventral surface of the head is labial in origin, be- cause the original bases of the second maxillz extended quite to the first pair of legs; an inconsiderable, if any, portion of the germ band inter- vening (Figure 21) between them. At Stage 8 the mouth is nearly closed (Plate 5, Figure 34) by the overgrowth of the combined second maxille and mouth-folds. In the adult (Plate 7, Figure 43) the apical lobes, although in con- tact mesally and stoutly chitinized, are readily separable and may be depressed and elevated by muscles homologous with those of Orchesella, the hinge lines being shown at swt. Shortly before hatching, hypodermis cells evaginate singly to form the external sete of the head. In the development of the labium, as I have traced it, neither galea nor lacinia becomes differentiated ; but the terminal lobe is equivalent to the head of the first maxilla, and therefore represents the common fundament of galea and lacinia, the second maxilla not passing the biramous stage. All of the labium behind the terminal lobe represents not only the stipes and cardo of the first maxilla, but also the mentum, submentum, and gula of the Orthopteran labium, —an important conclusion. In Orchesella (Folsom, ’99, Plate 3, Figure 24) mentum, submentum, and gula appear to be indicated, but the development in Anurida throws no light upon the structures which I suggested might be modified palpi. Packard (’71, p. 17) in describing Isotoma says, “I was unable at this or any other period to discover any traces of the second maxille. Though existing in a very rudimentary state in the adult, I could not detect them after repeated attempts, but do not doubt but that a more skilled observer would have made them out. Indeed, it is a most diffi- cult thing to discover their rudiments in the adult; I failed, at the time these observations were made, to detect them, though since then I have succeeded in making out their structure and relation to the surrounding parts of the mouth.” Asa matter of fact, he (Plate 3, Figure 13) has evidently figured the second maxille, which I know to be present in the genus, and in the passage quoted he doubtless referred to the super- lingue (“ paraglossze ”), which Lubbock also (’73, p. 66) termed “second maxillie.”’ Ryder, (’86, Plate XV. Figures 7, 9, 10), too, repeated the mistake in Anurida. Claypole (98, Plate XXIII. .Figures 40-44, 46, 47) represents the fundaments as simple papille without distinguishing the palpi, which are, however, obscurely indicated in Figures 43 and 47. FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 129 Uzel (’98, Taf. VI. Figur 87) shows the papillz of the second maxillze in Tomocerus. Concerning the development of the second maxille in Campodea, he (’98, p. 33) says, “Auch an den Anlagen der zweiten Maxillen (Taf. IV. Figur 38 und Taf. VI. Figur 79, mz.,) lasst sich ein kleinerer lateraler und ein grésserer medialer Theil unterscheiden, die indess nicht scharf von einander gesondert sind.” .. . (p. 34) “Anden Anlagen der zweiten Maxillen tritt auf der Mitte des Hinterrandes ein Vorsprung auf (Fig. 80, /e,), aus welchem sich, wie wir voraussenden wollen, der Lobus externus entwickelt, wogegen der friiher besprochene innere Theil den Lobus internus (/,) und der dussere den Palpus labialis (ymzx,) aus sich entstehen lasst. Zugleich bemerkt man an den beiden Maxillenpaaren eine gewisse Rotation. Die dusseren Enden derselben bewegen sich némlich nach vorn (Fig. 80), so dass die beiden Anlagen eine schriige Stellung erhalten. Bald jedoch, und zwar in dem Stadium, wo die vollkommene Umrollung des Keimstreifs zustande gekommen ist (Fig. 41), kehren sie in die urspriingliche Lage zuriick (das erste Maxillenpaare nicht ganz), und es erfolgt nun eine Rotation des zweiten Maxillenpaares allein im entgegengesetzten Sinne; die ansseren Enden desselben bewegen sich naémlich jetzt nach hinten und drehen die ganze Anlage in eine entsprechende schrage Lage, welche aus Fig. 81 ersichtlich ist. “Tm nachsten Stadium (Fig. 82) . . . die Anlagen des zweiten Max- illenpaares haben eine dreilappige Gestalt angenommen. Die drei Lap- pen lassen sich leicht deuten, wenn man die vorhergehenden Stadien vergleicht. Der vorderste (/,) entspricht dem Lobus internus, der mittlere (/e,) dem Lobus externus, und der hintere, breit gerundete (pmz,) stellt den Palpus labialis vor. Auch bemerken wir, dass sich nach der erwahnten Rotation die beiden Anlagen des zweiten Maxillen- paares einander stark in der Medianlinie genihert haben (Fig. 82) und auch etwas nach vorn geriickt sind.’ ‘In den nachsten Stadien (Fig. 83 und 84), bei welchem der Keimstreif schon etwas spiralig gerollt er- scheint (Fig. 42), riicken die Anlagen des zweiten Maxillenpaares noch naher aneinander, und zwar ganz besonders die Lobi interni (/¢).” In the postembryonic stage (p. 47): ‘Die beiden Anlagen des zweiten Maxillenpaares riicken in der Mittellinie noch néher als frither zusam- men, so dass nicht nur die Lobi interni (/iémz,), welche sehr gross geworden sind, sondern auch die Lobi externi (/emz,) dicht neben einander zu liegen kommen. Eine Verwachsung der beiden Halften des zweiten Maxillenpaares findet jedoch auch beim erwachsenen Thiere nicht statt.” 7 , 130 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. It is not clear, then, whether the galea develops from the outer or the inner lobe of a biramous appendage, although Uzel’s account is, at least, not inconsistent with his description of the first maxille, which I have already criticised. Although Uzel does not state as much, his figures indicate the palpus to be an appendage of the primary funda- ment, as it is in Anurida. In this genus, however, no third branch ap- pears, as I have said; but, from analogy with the first maxilla, the inner of the two branches represents undifferentiated galea and lacinia. The rotation in a frontal plane of the second maxillary fundament of | Campodea — which does not occur in Anurida— enables me to homo- logize the finished labium of Campodea with the apparently different labium of all other insects. If Uzel’s figures are compared with Figure 12 of vy. Stummer-Traunfels (91), it is easy to see that the embryonic structures by Uzel designated /imx, (lacinia), ema, (galea) and pmay (palpus) are with hardly a doubt respectively represented in the adult by the parts which vy. Stummer-Traunfels termed wp. (“untere Mundplatte”), pl. (‘‘tasterformige Papille”) and pp. (‘‘Tastwarze’’). These homo- logies, however, could never have been settled upon merely anatomical grounds. What Grassi (86°, Tav. IV. Figura 3), then, considered to be the under lip (da. 7m.) of Campodea is but the anterior part of the true labium; the “labial palpi” (pa. /.) are really galee borne upon a region representing the mentum, and the “labial papille” (pa. da.) are but modified palpi. As in Collembola, the labium is anteriorly and deeply cleft. Japyx is so close to Campodea that the same conclusions may doubt- less be applied to both genera. In Japyx the labium, as in Collembola, is split and bears a median sulcus (Grassi, 86°, Tav. III. Figura 21) much like that of Orchesella (Folsom, ’99, Plate 4, Figure 29). Ex- amining Figure 1 of v. Stummer-Traunfels (’91, Taf. IT.), the lacinia and galea are clearly represented, as in Campodea; the true palpus, however, is but obscurely differentiated in the region behind the so-called palpus (p/.) and nearer the median plane. The eversible papillz of the anterior part of the labium, as described by Meinert (’67, p. 369) and Grassi (’86, p- 31), are probably homologous with the papille of Orchesella which I designated plp. (99, Plate 3, Figure 24). For Lepisma, Heymons (97°, p. 590, Figur 11) gives, first, a pair of simple second maxillary fundaments and later (Taf. XXX., Figur 20) a long palpus with a small, basal, inner lobe, and states (p. 592) “ Die Lobi oder spi&teren Ladentheile der Maxillen sind in diesen Stadien erst FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA, ge | als sehr kleine unscheinbare Vorspriinge erkennbar, welche medialwarts an der Basis der Taster hervorwuchern.” This is contrary to the condi-’ tions in Anurida, where the palpus is certainly itself an outgrowth from the siraple, primary papilla (Figure 13). Lepisma agrees with Anurida, however, in that the galea and lacinia are derived from the inner lobe of a biramous fundament (Heymons,’ 97°, Taf. XXX. Figur 17), and dis- agrees with Campodea if, in the latter genus, as Uzel implies, the galea buds from the palpus. The finished labium of Lepisma, as I shall show, is remarkably like that of Orthoptera. The labium of Machilis, as described by Oudemans (’88, pp. 185-186, Taf. II. Figuren 28, 29), resembles that of Campodea and Collembola in being deeply cleft, and having the salivary ducts opening in similar posi- tions, but it more nearly approaches the Orthopteran type in the position and structure of the terminal lobes, the mentum, and the three-jointed palpi. Each terminal lobe is subdivided into four lobes, which in all probability collectively represent galea and lacinia. Ayers (84, p. 241, Plate 18, Figures 20-22; Plate 19, Figure 5), as already quoted (p. 124), has traced the development of the second maxille of Gicanthus as far as the trilobed stage, stating the lobation to be more prominent in the second than in the first maxillary appendage. The fact that the second maxille of Anurida develop upon the Orthopteran type is important. In Lepisma, the trilobed fundaments agree with those of Orthoptera even as to the greater length of the palpus. In the finished labium of Cicanthus (Packard, ’83*, Plate XXVII. Figure 9) the derivatives of each trilobed fundament are easily identified as three-jointed palpus, galea, lacinia, palpifer, and mentum, — the last two structures having doubtless arisen from the common stalk, or stipes. Although the labium is constructed upon the same plan in all Orthoptera, we may best select Blatta for comparison with Lepisma. The agreement between Blatta (Muhr, ’77, Taf. II. Figur 11; Packard, ’837, Plate XXVII. Figure 14) and Lepisma (Muhr, ’77, Taf. VIII. Figur 46; y. Stummer-Traunfels, ’91, Taf. II. Figur 17) is surprising. Gales and laciniz clearly correspond in the two, as do the mentum, palpifers, and palpi, the last, however, having three segments in Blatta and four in Lepisma. Muhr, in fact, included Lepisma among Orthoptera, as have some other authors. It is now agreed that the first and second maxille of Orthoptera are homodynamic, and, more inferentially, that the same is true of other in- sects. The exact agreement first recognized, according to Packard (’98, p- 69), by Miall and Denny (’86), was detected long before, at least, by 132 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Muhr (77, p. 9) and by Schaum (’61, p. 84). In Anurida the whole gular ~ region, excepting the terminal lobes and palpi, represents the undifferen- . tiated gula, submentum, mentum, and palpifers; therefore the gula in Orthoptera may be regarded as the united cardines, and the submentum, mentum, and palpifers, as stipal derivatives. It will be seen that my view differs from those accepted and defended by Packard (’98, p. 69) and — others; but it is supported by embryological evidence, while the other views are not. It may safely be predicted that the apparently unpaired gula of Orthoptera will be shown to originate from paired fundaments, as I have found it to do in Anurida. If these homologies between Collembola, Thysanura, and Orthoptera are accepted, their extension from the last group to other Pterygote orders is not difficult, even though the desirable embryological verifica- tions are still wanting. There is an unfortunate confusion of terminology regarding the mouth- parts of insects. The homologies are much obscured, but less by the use of different terms for homologous parts, than by the use of the same name for parts which are not homologous. “ Paraglosse” and “ligula” are cases in point. To most entomologists “ paraglossz ” mean indiffer- ently the labial lobes homodynamic with the galez and the lacinie of the first maxillze, or else mean the galeal lobes alone, while “ligula” or “ olossa’’ signifies the lacinial lobes, often more or less fused into a median organ; in fact, “ligula” is often used synonymously with “labium” in reference to many Coleoptera (Le Conte and Horn, ’83, p- xviii). “Ligula,’’ however, is often made a synonym of “lingua” (Packard, ’98, p. 68), and the latter term, of “hypopharynx.” In my opinion, the term “lingua” should be restricted to the median, un- paired constituent of the hypopharynx ; for the ‘‘ hypopharynx ” of cer- tain insects often bears two dorso-lateral lobes which in more generalized . insects are not only free from the lingua, but quite distinct from it in origin (as proved by myself in Anurida and by Uzel in Campodea), and these dorso-lateral appendages are most frequently called “ paraglosse,” upon assumptions which are not sustained by embryology, as I shall presently show. a As the terms “ paraglosse ” and “ ligula,” or “ glossa,” are irremoy- ably fixed, as applied to labial structures, they should not be used for anything else. It is both unnecessary and impossible to displace the term “hypopharynx,” but it is necessary to recognize the overlooked fact that the “ hypopharynx” is frequently a compound organ, to the ventral and median component of which the term “ lingua” may well FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 133 be restricted ; while, for the dorso-lateral appendages, rejecting ‘ para- gloss,” I propose the more appropriate name “ superlinguz.” The “gnathochilarium” of Symphyla and Diplopoda may also prove to be in part homologous with the hexapod labium. Having already discussed the resemblances between the lateral portions of the gnatho- ehilarium and the first maxille, I may compare the median components with the labium. They were, in fact, designated “under lip” in Sco- lopendrella by Grassi (’86*, Tav. II. Figura 5). As in Apterygota, there is a median portion and two stipal plates, each of which bears a papil- late head, separated by a transverse suture. These are the only points of agreement. On the contrary, the gnathochilarium is usually homologized with the first maxillz of insects (Packard, ’83°, p. 199; Korschelt u. Heider, ’90-93, p. 906) — apparently on account of Met- schnikoff’s (’74) researches. I can only suggest that the under lip of Diplopods is anatomically of too compound a nature to be homologized with the first maxillee only, and that we are not warranted in deriving the entire lip from only two primary fundaments simply because Met- schnikoff did not allude to more than two. In fact, Heymons (’97, p. 7, Figur 2) has discovered a “ post-maxillary”’ segment, without append- ages, in the embryo of Glomeris ; but he regards it as equivalent to the labial segment of insects. In other Diplopoda, for example Lysiopetalum “(Latzel, 84, Taf. IX. Figur 104) and Craspedosoma (Latzel, ’84, Taf. VL. Figur 72), the structure of the under lip is remarkably like that in Scolopendrella. In Chilopoda there are two fleshy, jointed appendages (“first maxilli- pedes,” “zweites Unterkieferpaar ”), which are conceivably equivalent, in position only, to the second maxille of Hexapoda, and are generally homologized with the first pair of legs of Diplopoda. If the second maxillz of insects are represented among Diplopoda in the manner I have suggested, then the second pair of Chilopod ‘ maxillipedes ” (“ Kieferfusspaar”’) corresponds with the first pair of feet of both Diplopoda and Hexapoda, —a simple conception. The labium of Hexapods is homologous with the first pair of maxilli- peds of Crustacea, according to the homologies which I have already proposed for all the more anterior paired appendages. It is, then, erro- neous to homologize with each other the second maxille in these two classes; but the error is so firmly established that I have in this paper frequently employed the term “second maxille” for the labium of insects, in order to avoid confusion. _ The evidence for my view of the homologies of the labium is of the VOL. XXXVI. — NO. 5 4 134 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. same character as that already used for the “ first maxille.” The labial fundaments are appendages of the seventh somite in both Hexapods and Crustacea and are supplied by equivalent ganglia and nerves. In both groups each fundament is at first simple and secondarily develops a— palpus, or exopodite. Moreover, the axis of the appendage is three- segmented, the segments in Crustacea corresponding to gula, mentum, - and palpifers of generalized Hexapoda, the submentum being a secondary development. Hansen (93, p. 206) differs slightly: “ Das Submentum [Machilis] ist mit dem, bei den Gammarinen zusammengeschmolzenen ersten Gliede — homolog, das Mentum mit dem bei den Hyperinen auch zusammenge- schmolzenen zweiten Gliede ; auf der Spitze des Mentums findet man ein — Glied, das auf jeder Seite in vier Laden ausgeht, die, wie sich ziemlich deutlich zeigt, zwei Laden angehdren, die jede fiir sich gespalten ist, und diese halte ich (unter Anderem wegen eines Vergleiches mit Orthoptera und Amphipoda, kann aber keinen zwingenden Beweis von den Skelet- theilen fiihren) respectiv fiir eine Lade vom zweiten Gliede (die innerste gespaltene Lade) und fiir das dritte Glied des Labiums mit seiner ge- spaltetem Lade; der Palpus geht von der Aussenseite des dritten Gliedes aus.” Hansen should have taken into consideration the gula, and the fact — that the submentum is probably not a primitive sclerite. The homologies between Hexapoda and Crustacea that I have defended are none the less valid if the total number of somites differs in the two classes, and they are sustained if the number is the same. In decapod Crustacea there are twenty-one somites, including the ocular segment. In generalized insects the number of abdominal segments varies. In the embryo of Lepisma, which shows marked affinities with Crustacea and Orthoptera, Heymons (’97*) finds eleven abdominal somites. Add to these the thoracic segments and the seven which I have found in the Apterygote head ; and the total, twenty-one, is the same as for decapod Crustacea. In embryos of many families of Orthoptera and Odonata just eleven abdominal segments are present. On the other hand, Heymons (’95°) has found twelve in certain genera of the same orders, — and in Collembola the number varies greatly. In view of this variability within the limits of the same order, then, it is well not to emphasize the — agreement between generalized insects and decapod Crustacea in the total — number of somites. . My conclusion regarding the labium, then, is that its development in~ Apterygota conforms to the Orthopteran type. In Anurida a labial pal — FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 135 _ pus is formed and resorbed, — an indication of degeneracy. The entire gular region of Apterygota is labial in origin; but fewer sclerites are differentiated than in Pterygote insects. The labium of insects is homo- dynamous with the “first maxillz ” and homologous in detail with the first maxillipeds of decapod Crustacea. The labium of Campodea is equivalent to the “second maxille” of Symphyla, and is represented in the Diplopod gnathochilarium. Skull. The principal mouth-parts of Collembola, unlike those of all other insects, except certain Thysanura, are internal ; the way in which they become so will now be described. The beginning of the process is seen at Stage 3 (Plate 1, Figure 3), ‘when the ventral surface of the germ band is almost flat. In lateral aspect (Plate 2, Figures 9, 10) the edge of the germ band is produced ‘downward as a small crescentic lobe (plz. or.) outside the fundaments of the mouth-parts. This lobe usually originates on the mandibular segment, as represented in Figure 9, but may arise more anteriorly, as in Figure 10. These figures represent, respectively, the left and right sides of the same individual. Rarely, the lobe begins behind the man- dible. A transection of the germ band near the middle of the lobe (Plate.3, Figure 16) proves the lobe (pli. or.) to be an evagination of the ectoderm enclosing mesoderm. In ventral aspect at this stage (Figure 11) the mouth-fold is clearly distinguishable at its widest part, or place of origin, but gradually disappears anteriorly and posteriorly on account of its confluence with the rest of the germ band. At Stage 4 (Plate 1, Figure 4) and a little later, while involution of the germ band is occurring, the mouth-fold is considerably larger (Fig- ures 12, 19, plz. or.) and forms a crescentic flap, now extending from the second maxilla almost to the labrum. In the next stage (Plate 1, Fig- ure 5) the fold is conspicuous ; in lateral aspect (Plate 3, Figure 20) its ventral margin is well rounded and conforms posteriorly to the con- tiguous anterior surface of the front leg ; the mandibular and maxillary fundaments still project slightly below the fold. In ventral aspect (Plate $, Figure 21) of tlie same individual, the fold is seen to be of nearly uniform width except anteriorly and posteriorly, where it is expanded against the labrum and second maxille respectively. Transections of the germ band (Plate 4, Figure 23), when compared with similar sec- tions at Stage 3 (Plate 3, Figure 16), show the folds to have exceeded 136 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. the mandibles in rate of downward growth, and the lateral surface of the mandible to be concave, in conformity with the swollen distal region — of the mouth-fold. In Stages 6 and 7 (Plate 1, Figure 6; Plate 2, Figure 7) the folds in- volve the labrum and second maxille (Plate 4, Figure 24 ; Plate 5, Figure — 30, pli. or.), covering the mandibles and first maxille laterally, and form- : ing the gene, or sides of the face. As seen in Stage 7 (Plate 5, Figure 30), each oral fold connects one side of the clypeo-labral fold with the labial evagination of the same side. There are no sutures, however, to — indicate the union of the gene dorsally with the clypeus and ventrally — with the second maxilla, for the oral evagination, in its backward and forward extension, has at length involved the labial and clypeal folds, respectively, in such a way that all three folds become one and enclose a single common cavity. The anterior margin of the mouth-fold is still — distinguishable, however, as late as Stages 7 and 8 (Plate 4, Figure 24, pli. or.) ; the mesal surfaces of the labial fundaments have not united anteriorly (Plate 5, Figure 29); the labrum is free from the fold (Figure ~ 30) and remains so. The mouth is definitely bounded, but still open (Figures 30, 34) ; its closure occurs, however, before the egg hatches. — The folds — clypeo-labral, oral, and labial — have been converging con-— comitantly with their elongation, and continue to elongate and converge until they meet to form a buccal cone, which completely encloses the — inner mouth-parts. After hatching, there is, for reasons just given, no demarcation of the mouth-fold ; it can simply be said that the region designated as pli. or. in Figure 40 (Plate 6) is the anterior part of that fold. Also in Orchesella the corresponding region, under which project the palpi (Folsom, ’99, Plate 2, Figure 9), doubtless originates as in Anurida, but the clypeus is not confluent with the folds. Strictly speaking, then, the mandibles and maxille are not ‘‘ retracted,” as is usually stated; but they are overgrown by the gene. Hansen (’93, p. 208) wrote concerning Campodea, Japyx, and Collem- bola, “die Mandibeln und Maxillen, mit Ausnahme der Spitzen, ‘im Kopfe liegen.’ Dieses ist dadurch entstanden, dass sich die Haut hinter ibrer Einlenkung wie eine Duplicatur, welche Gewebe enthalt, vorwirts und um sie herum gefaltet hat, und die Rander dieser Duplicatur sind auf der Unterseite des Kopfes mit dem Seitenriindern des Labiums festgewachsen, so dass dieses fast seiner ganzen Linge nach mit der Seitenwand des Kopfes verbunden ist.” These facts he ascertained by laborious dissections of the finished parts. ; Packard (’71, p. 21) simply mentions that “ the cephalic plates, which FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 137 fold back upon the head, forming the main expansion of the insectean head is [are] apparently the tergum of the antennary segment,” —a statement unsubstantiated by later and more extensive studies. The only account of the formation of the mouth-folds of Collembola is by Miss Claypole, who also studied Anurida maritima, giving her results briefly in 1896 and finally in 1898. The following extracts from her valuable paper (’98, pp. 264-266) summarize her observations and conclusions: “On each side of these [three pairs of mouth-parts, as in my Stage 3] has appeared a ridge that passes backward along the embryo, the two folds enclosing the mandibles and maxille. These folds start from just the region where the small intercalary appendages were seen earlier, but which have now disappeared. Figures 43, 46, and 47 show the process by which this change takes place, and leave no doubt that the folds, as they finally appear, are a development from the intercalary appendages. . . . The labrum in front and these lateral folds make together a three-sided box in which the mouth-parts, two mandibles, and four maxillz are sheltered. . . . The second pair of maxille has been _ modified to form the back of this pouch.” The author (pp. 265-266), after homologizing the neuromeres of Orthoptera and Crustacea, draws the important conclusion that the mouth-folds of Anurida “ including without doubt its allied forms,” are ‘clearly homologous with the second antenne ”’ of Crustacea. I quite disagree with this author as to the origin, and consequently the homology, of the mouth-folds. A priori arguments are here super- fluous, as the question is one of fact. As I have shown, the folds begin on, or very near, the mandibular segment, but always outside the paired fundaments of the mouth-parts, and never at the premandibular append- ages. The folds eventually and necessarily involve the intercalary region on progressing towards the labrum, although previously their early indi- cated continuity with the second maxillz (Plate 2, Figure 10) is estab- lished. Conceptions as to the development of the fold are of course but inferences from facts observed in certain stages. The most apparent inference from the figures cited by Miss Claypole as leaving no doubt about the accuracy of her conclusion is certainly the one she has drawn ; but from the same figures and from her preparations — which Miss Claypole has most kindly lent me— may also be drawn the less ap- parent, though I believe correct, inference that the folds begin between the interealary and second maxillary regions and grow towards both of them. I have found stages intermediate between those shown by Miss Claypole in Figures 46 and 47, which convince me that this is the ae) 138 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. case. Consequently the mouth-folds cannot represent the Crustacean second antennz. My own views as to the homology of the mouth-folds, — already implied by my use of the term “gene,” will presently be supported. Hansen’s recognition of the similarity between Campodea, Japyx, and Collembola is sustained by embryology. In Campodea, Uzel (’98, p. 33) describes and figures a “ Chitinstrang . . ., welcher sich von der Vorder- randmitte der zweiten Maxille um die Aussenseite des ersten Maxille — und der Mandibel herum zu den auf den Intercalarsegmente gelegenen Hockern zieht.” His Figures 38 and 79 show the Chitinstrang at a — rather advanced stage of development, corresponding with the condition in my Figure 17; unfortunately he gives neither its origin nor its earlier development. The later development, as evidenced diagrammatically by — his Figures 80-84, agrees with that of Anurida in the gradual approxi- | mation of the lateral ridges, and especially in the completion of the buccal boundary by the same method of confluence. Uzel does not attempt to explain the homology of the Chitinstrang. In Lepisma and Machilis the mouth-parts are ectognathous, as in Orthoptera. In Lepisma there is no trace of a lateral mouth-fold, but in Machilis I have found a distinct, flat, lateral lobe sheltering the base of each mandible, and the lobe is probably homologous with the Collem- — bolan mouth-fold. In Pterygota the gen, often not demarcated as distinct sclerites, represent the lateral regions of the germ band —as they do in Campodea — and Collembola. In these Apterygota the same areas have simply in- creased as folds, but the folds are none the less homologous with the pleural regions of other insects, and in Collembola are reasonably to be regarded as the pleural portions of the premandibular, mandibular, and both maxillary segments. In many Pterygote insects, especially in Orthoptera, the gen overlap the bases of the jaws; for example, in — Caloptenus, in which the gena is produced as a small but distinct flat fold over the base of the mandible. Little is known about the development of the sides of the head in Myriopoda, but in Peripatus it is interesting to find distinct lateral mouth-folds (Sedgwick, ’88, Plate IT., Figure 36) quite analogous, to say the least, with those of Collembola. / Concerning the completion of the skull, little remains to be said. At Stage 7 a constriction encircling the blastoderm separates the head from — the thorax. The head is typically a hollow cylinder, or cone, and so is the body. The body cylinder consists of a definite number of successive | FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA, 139 rings, in each of which, in the more specialized insects, tergum, pleura, and sternum are present. In the head region of the Collembola, however, segmentation occurs only on the ventral side of the germ band. The entire gular region is Jabial in origin, and there is reason for regarding the clypeus as the tergite of the ocular segment. The mouth-folds are undoubtedly ex- panded pleura. Aside from these, however, it is idle to speculate about the location of other sclerites which are differentiated in more spe- cialized insects. Here, in the absence of such differentiation, it may be be said that the head-cylinder represents seven ideal rings, which dorsally and laterally are in no way demarcated from each other. Admitting that the procephalic lobes do extend backward and encroach upon other segments, the lobes may not be regarded as the tergites or pleurites of those segments, for they are simply thickened blastoderm, and increase in area in proportion as the blastoderm thickens; but the convenience of applying a single term, ‘ procephalic lobe,” to either of these thicken- ings should not blind us to the fact that the lobe eventually represents the blastoderm of more than one segment. In the finished head (Plate 5, Figure 33) are certain elevated dorsal areas which, however, are not sclerites bounded by sutures, and are not clearly to be homologized with sclerites of other hexapod orders. The elevations referred to are directly correlated with underlying glands and muscles. The sides of the face in Apterygota, then, are homologous with the gene of Pterygota. Im all insects the skull represents seven somites, but the cephalic sclerites of Pterygota, excepting labrum, clypeus, and labial sclerites, are not differentiated in the Apterygota. Tentorium. The tentorium of Anurida is essentially like that of Orchesella (Folsom, *99), consisting of a chitinous plate parallel with the frontal plane (Plate 8, Figure 51, tnt.), from which diverge two pairs of chitinous arms (Plate 6, Figure 35) extending to the skull: a dorsal pair (dr. d.) and a poste- rior pair (dr. p.) embracing respectively the supra- and infra-cesophageal ganglion. A third, or anterior, pair was found in Orchesella, but not in Anurida. Regarding the development of the tentorium in insects, most diverse Opinions are held. After considerable study, I have come to the con- 140 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. clusion that the tentorium of Anurida is derived from proliferated ecto- dermal cells which are in no way, except in position, distinguishable from young ganglion cells. In Anurida, as in Orthoptera (Wheeler, ’93, Heymons, ’95>) and Lepisma (Heymons, 97°), the ventral cords consist of dorso-ventral rows of cells, which arise by proliferation from the outer ectoderm, Although it has seldom been supposed that these cells became other than ganglionic in function, it may be assumed, in view of their origin, that all of them are potentially chitin-forming cells, and it seems prob- able that some of them actually do form the chitinous tentorium. An oblique section of Stage 8, cut at a fortunate angle for studying the relation of tentorium to cells, gave the appearance represented in Figure 35. Contiguous to practically all parts of the tentorium, in this section, are cells the nuclei of which do not differ in appearance from nuclei of undoubted ganglion cells. On all sides of the tentorium such cells abound and closely embrace it ; an especially large mass of these cells occurs immediately under the frontal plate, in which, moreover, several cells always become enclosed and appear to be functional in the adult. I found no evidence which could be interpreted as indicating any other way of formation. Von Stummer-Traunfels (91) appears to have overlooked the ten- torium of Apterygota, for he mentions the “Stiitzapparate” only, by which he evidently means the structures I call “lingual stalks.” As regards the Thysanura, Meinert (’65, Tab. XIV. Figur 5, 0) men- tions in Japyx and Campodea a median chitinous plate, from which the mandibular adductors take their origin, which is undoubtedly the tento- rium. Grassi (86”) also alludes to it in Japyx. In Machilis the lingual stalks, important in Collembola, become rudi- mentary ; and most of the mandibular and maxillary muscles become attached to the tentorium ; but they are fewer than in Collembola, The tentorium is thus described by Oudemans (’88, p. 186): ‘ Die vor- deren [Stiitzplatten] kommen von den Seiten des Clypeus, gleich ober- halb der Mahlhdhle, wie dieses im Durchschnitt abgebildet ist in Figur 32. Links und rechts geht dort die Chitinhaut des Clypeus tiber in eine Platte. Die beiden Platten niihern sich nach hinten, indem sie fortwiih- rend breiter werden. In der Mitte des Kopfes kommen sie zusammen, sind da jedoch nicht verschmolzen, sondern nur stark durch Bindegewebe verbunden, Figur 19 L’. Hinter dieser Verbindungsstelle weichen die Platten wieder auseinander, werden schmiiler und gehen, links und — rechts vom (Esophagus, nach oben. Zuletzt geht jede tiber in einen Cali, ell et a a ei FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 141 diimnen Chitinstab, welcher oben im Kopfe, hinter den Augen, endet und da am Chitin des Kopfes festsitzt.” Thus, the tentorium of Machilis is constructed upon the same plan as that of Anurida, although the median plate is halved longitudinally. The dorsal and posterior arms in Anurida are clearly represented in Machilis, and the latter pair tends to become reduced in size, — an approach to the Orthopteran condition. The tentorium in Orthoptera is readily comparable with that of Machilis. In Periplaneta, according to Miall and Denny (’86, p. 39), “Tn front it gives off two long crura, or props, which pass to the gin- glymus, and are reflected thence upon the inner surface of the clypeus, ascending as high as the antennary socket, round which they form a kind of rim. Each crus is twisted, so that the front surface becomes first internal and then posterior as it passes towards the clypeus. The form of the tentorium is in other respects readily understood from the figure (Figure 17). Its lower surface is strengthened by a median keel which gives attachment to muscles. The wsophagus passes upwards between the anterior crura, the long flexor of the mandible lies on each side of the central plate ; the supracesophageal ganglion rests on the plate above, and the subcesophageal ganglion lies below it, the nerve- cords which unite the two passing through the circular aperture. A similar internal chitinous skeleton occurs in the heads of other Orthop- tera, as well as in Neuroptera and Lepidoptera.” In Anabrus (Packard, ’98, p. 49, Figure 33) the tentorium is essen- tially the same, with a central plate, and paired dorsal and posterior arms. The only important differences between Orthoptera and Collem- bola in respect to the tentorium are (1) that the esophageal commissures pass through it in the former group instead of around it; (2) that in Orthoptera the posterior arms are reduced in length, and (3) that the tentorium becomes more stoutly chitinized. On the other hand, the tentorium of Orthoptera, in its general form and topographical relations, agrees closely with the same structure in Collembola and Thysanura. Palmén (’77) derived the tentorium from a pair of cephalic trachee in Ephemera, but upon unsatisfactory grounds. In Collembola trachez are absent ; moreover, as Packard (’98, p. 50) notes, ‘the apodemes of the thoracic region are evidently not modified trachez, since the stigmata and tracheze are present.” The views of Carriére (90) and Cholodkowsky (91), agreeing with the opinion of Palmén, have been controverted by Heymons (’95, pp. 50-51), 142 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Wheeler (’89, p. 568) finds that five pairs of ectodermal invaginations form the tentorium of the larval head of Doryphora. ‘These invagina- tions grow inwards as slender tubes, which anastomose in some places. Their lumina are ultimately filled with chitin.” Wheeler offers his observations in support of Palmén’s theory, but they are not at all inharmonious with the scanty observations I have made upon Anurida. Heymons ('95°, pp. 50-51), describing Forficula, agrees with Wheeler, except that he finds only two pairs of fundaments for the tentorium, and says (p-51): “Ich habe mich indessen davon iiberzeugt, dass auch bei Gryllus und Periplaneta die zahl der Tentoriumanlagen keine gréssere ist, sondern, wie Heider (’89) dies bei Hydrophilus beschrieb, und ich es bei Forficula fand, nur vier betragt. Der oben geschilderte Entwicke- lungsmodus des Tentoriums diirfte daher wohl als der typische anzusehen sein.” In Anurida I was unable to find any distinct ectodermal invaginations which might form the arms of the tentorium, but am not prepared to say that none exist, because the subject is one of great difficulty. The arms must be studied in oblique sections, and it is almost impossible to dis- tinguish them from fundaments of muscles until they are nearly com- pleted. The finished tentorium of Collembola, however, is undoubtedly homologous with that of Thysanura, and almost as clearly with the ten- torium of Orthoptera. Segmentation of the Head. The elucidation of the primitive segments in Arthropods is a most interesting and difficult morphological problem. The rule of Savigny, — emphasized by Huxley and others, — that Arthropods consist funda- mentally of successive rings, each of which may bear but one pair of primary appendages, although now undoubted, has never been thoroughly substantiated when applied to the Hexapod head. After vears of argu- ment, morphologists still disagree as to the number of somites composing the highly differentiated heads of insects. Kolbe (’90, p. 135) recognizes five, as follows : — 1. Ursegment: Fithler, Augen, Oberlippe ; . Ursegment : Oberkiefer oder Mandibeln (1. Kiefernpaar) ; . Ursegment: Unterkiefer oder Maxillen (2. Kiefernpaar) ; . Ursegment : Zunge oder Innenlippe (3. Kiefernpaar, verwachsen) ;— . Ursegment: Unterlippe (4. Kiefernpaar, verwachsen). Sharp (95, p. 87) says, “ Morphologists are not yet agreed as to their Or em © bo | FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA 143 number, some thinking this is three, while others place it as high as seven ; three or four being, perhaps, the fgures at present most in favor, though Viallanes, who has recently discussed the subject, considers six, the number suggested by Huxley, as the most probable. is of a similar opinion.” Packard (’98, p. 54) gives six :— Pre-oral in early embryo. Post-oral in early embryo. NAME OR SEGMENT. . Ocellar (Protocerebral). . Antennal (Deutocerebral). . Premandibular, or Intercalary ( Tritocerebral). . Mandibular. 5. First Maxillary. . Second Maxillary, or labial. PrecEs oR REGIONS OF THE HEAD-CAPSULE. Epicranium, ante- rior region with the clypeus, la- brum, and epi- pharynx. Epicranium, includ- ing the antennal sockets. Wanting in post- embryonic life, ex- cept in Campodea. Epicranium behind the antenne, gene. Epicranium, hinder edge? tentorium. Occiput. Cholodkowsky APPENDAGES, ETC. Compound and sim- ple eyes (ocelli). Antenne. Premandibular appendages (in Campodea). Mandibles. First maxille. Second maxille, or labium. Post- gula, gula, sub- mentum, mentum, hypopharynx (lin- gua, ligula), para- glosse, spinneret. Upon anatomical grounds, different observers have recognized from one to seven head segments. As mentioned by Packard (’98, p. 50), Burmeister found but two; Carus and Audouin three ; MacLeay, New- man, and Newport four ; Straus-Durckheim seven. Huxley (’78, p. 343) said: “It is hardly open to doubt that the mandibles, the maxille, and the labium answer to the mandibles and the two pairs of maxillz of the crustacean mouth. In this case, one pair of antennary organs found in the latter is wanting in insects, as in other air-breathing Arthropods, and the existence of the corresponding somite cannot be proved. But if it be supposed to be present, though without any appendage, and if the 144 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. eyes be taken to represent the appendages of another somite, the insect- head will contain six somites.””. . . Huxley’s conclusions were the most satisfactory that could be derived from a study of the completed organs alone, and reduced the problem to these questions: Do the eyes represent a somite? Is another antennal segment represented in insects? Do the labrum and hypopharynx repre- sent distinct segments ? Authors began to realize the impossibility of settling the problem upon purely anatomical data, and attacked it from the embryological side. Packard (71, p. 21) concluded, ‘“ Accordingly, we seem forced to the belief that the head of the hexapodous insects consists of but four seg- ments, 7. e. the second maxillary, first maxillary, and mandibular seg- ments, situated behind the mouth opening, and the antennary, or first and pre-oral segment, situated in front of the mouth. . . . The clypeus and labrum are apparently differentiated from the cephalic lobes, and thus seem to form a portion, or fold, of the antennary segment.” Graber (’79) reached the same conclusion. Viallanes, after carefully studying the development of the nervous system in Insects and decapod Crustacea, wrote the most important contribution upon the subject that has yet been published, and gave his results as follows (87, pp. 108-109) :— “1. Le cerveau des Insectes, comme le cerveau des Crustacés décapodes, est formé de trois segments : j’appelle le premier protocerebron (cerveau du pre- mier zoonite) ; le deuxiéme, deutocérébron (cerveau du deuziéme zoonite) ; le troisiéme, tritocérébron (cerveau du troisiéme zoonite). “2. Nous retrouvons, dans Je protocérébron de |’Insecte, toutes les parties constitutives du protocérébron du Crustacé. Dans cette premitre région céré- brale, la seule différence qui s’observe entre les deux types est la snivante ; chez V'Insecte les deux lobes protocérébraux viennent se souder sur la ligne médiane et se mettre ainsi en contact avec le protocérébron moyen. Chez le Crustacé, au contraire, les lobes protocérébraux sont tres écartés de la ligne médiane et logés dans les pédoncules oculiféres. Le rapprochement qui, chez l’Insecte, s’effectue entre les lobes protocérébraux, entraine la disparition, ou pour mieux dire le raccourcissement extréme du tractus nerveux connu chez les Crustacés sous le nom de nerf optique. “3. La deutocérébron, qui a une structure extrémement caractéristique, se retrouve chez |’Insecte et chez le Crustacé avec les mémes caractéres et les mémes connexions. [1] en résulte que le nerf antennaire de I'Insecte est Vhomologue du nerf de l’antennule du Crustacé. “4. Chez le Crustacé le tritocérébron se compose des deux lobes antennaires et des deux ganglions @sophagiens (improprement appelés mandibulaires) et d'une commissure sous-cesophagienne (la commissure transverse de ’anneau FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 145 ceesophagien) qui réunit ces derniers. Le lobe antennaire donne naissance au nerf de l’antenne externe, le ganglion cesophagien 4 Ja racine du premier gan- glion visceral impair (ganglion stomatogastrique) et au nerf du labre. “ Chez l’Insecte, le tritocérébron subit une importante réduction, les lobes et les nerfs antennaires disparaissent, mais les représentants des ganglions esophagiens (que j’ai décrits sous le nom des lobes tritocérébraux) subsistent dans leur intégrité. Comme chez le Crustacé, ils donnent naissance 4 la racine du premier ganglion viscéral impair (ganglion frontal) et au nerf du labre ; comme chez les Crustacés, ils sont unis l’un & Vautre par une commissure sous-cesophagienne (commissure transverse de l’anneau cesophagien). Ainsi: - “Le nerf de l’antenne externe du Crustacé n’est pas représenté chez l’Insecte. “Qe nerf du labre de V’Insecte est homologue du nerf du Jlabre du Crustacé.” After a lengthy discussion of the segmentation of the head, Viallanes concludes (87, pp. 117-118) : — “1. La téte de l’Insecte est formée par six zoonites, trois sont pré-buccaux et trois post-buccaux. “9. Le premier zoonite porte les yeux composés et les ocelles. Le deuxieme les antennes. Le troisiéme, qui est dépourvu d’ appendices, porte le labre, piece qui, pas plus chez les Insectes que chez les Crustacés, ne peut étre considérée comme le résultat de la soudure de deux appendices. «3. Le quatriéme zoonite porte les mandibules, le cinquiéme les machoires, le sixiéme la lévre inférieure.” Wheeler (93), Heymons (’95*), and others have confirmed these conclusions. Heymons (’95°, p. 36), in a valuable paper on the segmentation of the insect-body, says, “ Der Kopf besteht aus sechs Korperabschnitten : dem Oralstiick, Antennensegment, Vorkiefersegment, und drei Kieferseg- ‘menten.” Rudimentary intercalary appendages have been found in Anurida (Wheeler, ’93) and Campodea (Uzel, 97°). Claypole (’98) and Uzel (’98) have homologized them with the second antenne of Crustacea. Six somites are the most that have been admitted upon embryological grounds, but I am convinced that there are more than six. Hansen (’93) suggested that the so-called ‘“‘ paraglosse ” [superlinguze] of Machilis were homologous with the Crustacean first maxilla, and my observations upon the development of the superlinguz support his view. The superlingue originate independently as a pair of simple papille — like the mandibles and maxilla — intermediate between the mandibles and the “ first maxille,” and represent a distinct, though reduced, seg- 146 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. ment, because provided with a ganglion. More conclusive proof could hardly be expected. The insect-head, then, is composed of seven somites, which are homolo- gous with the first seven of decapod Crustacea. If the conclusions I have drawn in this paper are valid, certain radical changes become necessary in the commonly accepted ideas of homology among the great classes of Arthropods. These changes I submit in the following table : — TABLE OF EQUIVALENT SOMITES IN THE HEAD OF ARTHROPODA. Segment | Arachnida | Chilopoda ; Hexapoda Compound | Compound eyes and eyes and ocelli ocelli Embryonic “4 A BABE First Antenne preantenne. antenne Antenne Antenne Second Intercalary antenne appendages Chelicere | Mandibles Mandibles Mandibles Mandibles Pedipalpi First First Superlingue maxille maxille First legs Second [Gnathochilarium Second Maxille maxille maxille Second legs Maxillipedes First Labium maxillipedes Summary. The protocerebrum of Apterygota agrees with that of other insects in development and structure. The ocular segments of Hexapoda and — decapod Crustacea, as well as the compound eyes of the two groups, are ~ homologous. The labrum and clypeus of insects develop from a single median evagination between the procephalic lobes, and do not represent a pair of appendages. The labrum of Apterygota is homologous with that of other insects, as well as that of Symphyla, Diplopoda, Chilopoda, and the higher Crustacea, FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA, 147 The antenne of Apterygota evaginate from the posterior boundaries of the procephalic lobes, and therefore agree with those of Pterygota in this respect. In both groups the antennz are at first post-oral and sub- sequently pre-oral in position. The deutocerebrum of insects is homologous with that of Crustacea, and the antenne of Hexapoda are equivalent to the antennules of Crustacea and the embryonic preantenne of Chilopoda. Premandibular, or intercalary, appendages exist in the embryo of Anurida, and appear to be represented even in the adults of several Apterygote genera. The tritocerebrum of Apterygota is homologous with that of Orthoptera and decapod Crustacea, and the rudimentary premandibular appendages of Collembola and Thysanura represent the second antennze of decapod Crustacea and probably the antenne of Diplopoda and Chilopoda. A distinct primitive ganglion occurs in the intercalary segment of Anurida, therefore the segment must be regarded as one of the primary head-segments. The mandibles of Apterygota develop from a pair of simple papille, the bases of which become oblique. No trace of lobation occurs except in Campodea. The mandibles of Collembola and Thysanura are homo- dynamous with the maxille and homologous with the mandibles of Pterygota, Scolopendrella, Crustacea, and probably Diplopoda and Chilopoda. The “hypopharynx” in Apterygota is a compound structure consist- ing of two dorsal “superlingue,” —as I have called them, — which develop from a pair of papillee between the mandibular and first maxil- lary segments, and also a ventral lingua, which originates independently as a median unpaired evagination on the first maxillary segment. The two chitinous “lingual stalks,” which are most highly developed among Apterygota, arise in superficial grooves of the ectoderm. The hypo- pharynx of Apterygota is undoubtedly homologous with that of Ptery- gota; although, in the latter group, the lingua and superlinguze become united together and the lingual stalks become rudimentary. In Anurida a distinct neuromere exists for the “superlingue ;” therefore it is neces- Sary to recognize the superlingual segment as equivalent in morphologi- cal value to the other primary somites. The superlingue are homologous with the first maxillze of Malacostraca and Chilopoda and are anatomi- cally represented in the labial plate of Diplopoda. In order to avoid confusion, the terms “ paraglossee” and “ligula” should not be applied to the constituents of the hypopharynx, but are better restricted to the labium of insects. The lingua of Hexapoda is equivalent to the Crusta- 148 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. cean hypopharynx, and possibly also to the median component of the Diplopod gnathochilarium. The first maxillee in Collembola and Thysanura develop essentially as in Orthoptera and may be homologized part for part with the maxille of generalized Pterygota. In Anurida a palpus appears in the embryo, but is resorbed before hatching, indicating the derivation of this genus from a form in which the first maxillary palpi were functional, as they are at present in Orchesella, Tomocerus, and other Collembolan genera. The first maxillee of Campodea are clearly to be homologized with the first of Scolopendrella, the second of Chilopoda, and less clearly with the — lateral portions of the Diplopod gnathochilarium. The first maxille of insects pass through a biramous condition, as in Crustacea, and the sclerites of these organs appear to be homologous in the two groups; the first maxille of Hexapoda, however, are equivalent to the second maxillz of Malacostraca. The labium in Anurida develops from a pair of papille, from which the entire gular region is derived. A palpus appears, but is soon re- sorbed, and no galeal and lacinial lobes are differentiated. Upon the whole, the labium among Apterygota is homologous with the same structure of Pterygota, although fewer sclerites are formed in the former group. The labium in insects, homodynamous with the man- dibles and first maxille, agrees in detail with the first maxillipedes of — decapod Crustacea. The labium of Campodea is homologous with the “second maxille” of Scolopendrella and the maxillipeds of Chilteots and is represented in the gnathochilarinm of Diplopoda. The sides of the face in Anurida develop from two lateral evaginations of the germ band near the mandibular segment, which eventually involve the labral and labial fundaments and complete the buccal cone. The mouth-folds of Collembola, Campodea, and Japyx are strictly homologous with the gene of Pterygota. The dorsal region of the skull in Anurida does not differentiate into sclerites which may be compared with those of Pterygote insects. The tentorium is inferred to develop from cells which have been pro- liferated from the ectoderm. The evidence convinces me that there are just seven somites in the head of Anurida, and that probably the same is true for all Hexapoda, The cephalic somites are successively: ocular, antennal, intercalary, — mandibular, superlingual, maxillary, and labial. As I have found embryonic ganglia for the intercalary and superlingual segments, there are seven cephalic ganglia, one for each somite. Moreover, excepting FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA, 149 the ocular segment, every somite is represented by a pair of append- ages. I find no evidence whatever for more than seven primitive cephalic segments, and believe that my observations have assisted to settle the long-disputed question of the segmentation of the head. Since the time of Fabricius, the mouth-parts of insects have been of primary importance for the systematist. While insisting that a logical classification must recognize all anatomical structures, it must be ad- mitted that the mouth-parts are of fundamental systematic value on account of the range of their differentiation. ; Without discussing at length the phylogeny of insects, I may briefly give the bearing of these studies upon the subject, remarking that my conclusions are in entire accord with approved views upon the origin of insects. The Collembola are strikingly like Campodea and Japyx in structure, their peculiar entognathous characteristic separating these three groups from all other insects. The Collembola as a group are somewhat more specialized than the Thysanura in general structure. The Smynthuride, with their globular bodies, vertical heads, and well-developed furculz and ventral tubes, represent one extreme of differentiation — compara- tively high. The Aphoruride, including Anurida, with vermiform bodies, subequal segments, horizontal heads, no furcula, ete., are much more generalized, and probably degenerate forms. Anurida, for ex- ample, has both pairs of maxillary palpi, as well as rudimentary ab- dominal appendages and the fundaments of a furcnla in the embryo, but in the embryo only. Therefore the ancestral Collembolan was probably intermediate between Smynthuride and Aphorurid, and is most nearly represented by members of the family Poduride. The resemblance in the mouth-parts leads us to suppose that the primitive Collembolan descended from the stem form of Campodea and not far below Campodea itself. The affinities of Campodea, which is slightly more primitive than Japyx, are in two directions: towards Machilis and Lepisma on the one hand, and towards Scolopendrella on the other. In the first two genera the mouth-parts are clearly derivable from the Campodean type, and link Campodea with Orthoptera. In regard to Scolopendrella, it was long uncertain whether it should be placed among Thysanura or Myriopoda, on account of its strong affinities for both. Most authors have followed Grassi and placed it in the latter group, always admitting its insectean features. In the mouth-parts, Scolopendrella approaches VOL. XxXVI.— No. 5 5 150 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Campodea rather than Diplopoda, but is unquestionably nearer Diplo- poda than it is to Chilopoda. The gnathochilarium of Diplopoda may be homologized with the appendages of three hexapod somites, but only two embryonic segments have as yet been found ; and the subject needs further investigation. The mouth-parts of Chilopoda may be homologized with those of insects in only the broadest way, the correspondences being principally those of position. Between decapod Crustacea and Apterygota there are decided mor- phological resemblances. The seven cephalic somites which I have found in the latter group I have homologized in detail with the anterior seven of the former, and pointed out that most of the homologous ap- pendages function alike in the two groups. These homologies, however, simply indicate a partial parallelism in development; for in most re- spects Crustacea and Hexapoda are very divergent classes. LT Els 5 tn la ae cd orr> FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA, 151 BEBLLIOGHRAPHY, Ayers, H. 84. On the Development of (canthus niveus and its Parasite, Teleas. Mem. Bost. Soc. Nat. Hist., Vol. 3, No. 8, pp. 225-281, Pl. 18-25, 41 fig. Burgess, E. *80. Contributions to the Anatomy of the Milk-weed Butterfly (Danais ar- chippus Fabr.). Anniv. Mem. Bost. Soc. Nat. Hist., 16 pp., 2 pl. Biitschli, O. 70. Zur Entwicklungsgeschichte der Biene. Zeits. f. wiss. Zool., Bd. 20, pp- 519-564, Taf. 24-27. Carri€re, J. 790. Die Entwicklung der Mauerbiene (Chalicodoma muraria Fabr.) im Ei. Arch. f. mikr. Anat., Bd. 35, pp. 141-165, Taf. 8, 8a. Cholodkowsky, N. 91 =Die Embryonalentwicklung von Phyllodromia (Blatta) germanica. Mém. Acad. Imp. Sci. St. Pétersbourg, Sér. 7, T. 38, (4), 120 pp., 6 Taf. Translation of § 3 (pp. 86-101): On the Morphology and Phylogeny of Insects. Aun. Mag. Nat. Hist., Ser. 6, Vol. 10, pp. 429-451. 1892. Claypole, A. M. "96. The Appendages of an Insect Embryo. Can. Entom., Vol. 28, p. 289. [Report by D. S. Kellicott, Secy. Sec. F., Amer. Assoc. Adv. Sci.] Claypole, A. M. *98 The Embryology and Oégenesis of Anurida maritima (Guér.). Jour. of Morph., Vol. 14, pp. 219-300, Pl. 20-25, 11 fig. Dimmock, G. 81. The Anatomy of the Mouth-Parts and of the Sucking Apparatus of Some Diptera. 50 pp., 4 pl. Boston. Dugés, A. *32. Recherches sur les caractéres zoologiques du genre Pulex, et sur la Multiplicité des espéces qu'il renferme. Ann. Sci. Nat., Zool., T. 27, pp. 145-165, Pl. 4. Pe 152 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Folsom, J. W. °99. The Anatomy and Physiology of the Mouth-Parts of the Collembolan, Orchesella cincta L. Bull. Mus. Comp. Zodl., Vol. 35, No. 2, pp. 7-39, 4 pl. Grassi, B. 85. Studi sugli Artropodi. Intorno allo sviluppo delle api nell’ uovo. Atti Accad. Gioen. Sci. Nat. Catania, Ser. 3, T. 18, pp. 145-222, 10 tav. Grassi, B. °86a. I progenitori degli Insetti e dei Miriapodi. [Memoria I.] Morfologia delle Scolopendrelle. Mem. Reale Accad. Sci. Torino, Ser. 2, T. 37, pp- 593-624, 2 tav. : Grassi, B. 86>. I progenitori degli Insetti e dei Miriapodi. [Memoria II.] L’Japyx e la Campodea. Atti Accad. Gioen. Sci. Nat. Catania, Ser. 3, T. 19, pp. 1- 83, Tav. 1-4. Grassi, B. ’86¢. I progenitori dei Miriapodi e degli Insetti. Memoria III. Contribu- zione allo studio dell’ Anatomia del genere Machilis. Atti Aecad. Gioen. Sci. Nat. Catania, Ser. 3, T. 19, pp. 101-128, 1 tay. Hansen, H. J. °93. Zur Morphologie der Gliedmassen und Mundtheile bei Crustaceen und Insecien. Zool. Anz., Jhg. 16, pp. 193-198, 201-212. Hansen, H. J. 94. On the Structure and Habits of Hemimerus talpoides Walk. Entom. Tidsk., Arg. 15, pp. 65-93, Pl. 2, 3. Heider, K. °89. Die Embryonalentwicklung von Hydrophilus piceus L. 98 pp., 13 Taf. Jena. Heymons, R. 95a, Die Segmentirung des Insectenkérpers. Anhang z. d. Abh. Kongl. Preuss. Akad. Wiss., Berlin, 39 pp., 1 Taf. Heymons, R. ’°95b. Die Embryonalentwickelung von Dermapteren und Orthopteren unter besonderer Beriicksichtigung der Keimblatterbildung. viii, 136 pp., 12 Taf. u. 33 Fig. Jena. Heymons, R. 96. Grundziige der Entwickelung und des Kérperbaues von Odonaten und Ephemeriden. Anhang z. d. Abh. Kéngl. Preuss. Akad. Wiss., Berlin. 66 pp-, 2 Taf. Heymons, R. 97a, Entwicklungsgeschichtliche Untersuchungen an Lepisma saccharina L. Zeits. f. wiss. Zool., Bd. 62, pp. 583-631, Taf. 29, 30. 3 Fig. Ps ° FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 153 Heymons, R. '97>. Mittheilungen iiber die Segmentirung und den K6érperbau der Myrio- poden. Sitzb. Kéngl. Preuss. Akad. Wiss., Berlin, Bd. 40. pp. 915-923. 2 Fig. Separate, 9 pp-, 2 Fig. Hollis, W. A. *72. The Homologue of a Mandibular Palp in Certain Insects. Jour. Anat. Phys., Vol. 6, pp. 895-897, Pl. 18. Huxley, T. H. 78. A Manual of the Anatomy of Invertebrated Animals. 596 pp., 158 fig. New York. Kirby, W., and Spence, W. 28. An Introduction to Entomology. Ed. 5, 4 vols. London. Kolbe, H. J. 89-93. Linfihrung in die Kenntnis der Insckten. xii, vill, 709 pp., 324 Fig. Berlin. Korotneff, A. 85. Die Embryologie der Gryllotalpa. Zeits. f. wiss. Zool., Bd. 41, pp. 579- 604, Taf. 29-381. Korschelt, E., und Heider, K. 90-93. Lehrbuch der vergleichenden Entwicklungsgeschichte der wirbel- losen Thiere. xii, 1509 pp., 899 Fig. Jena. Latzel, R. 84. Die Myriopoden der dsterreichisch-ungarischen Monarchie. Zweite Halfte. xii, 414 pp., 16 Taf. Wien. Le Conte, J. L., and Horn, G. el 83. Classification of the Coleoptera of North America. Smiths. Miscel. Coll. [No.] 507. xxxviii, 567 pp. Washington. Lemoine, V. 83. Recherches sur le développement des Podurelles. Assoc. Praga: Sci., 1le Session (1882). pp. 483-520. Pl. 14-16. Lubbock, J. f 73. Monograph of the Collembola and Thysanura. Ray Soc. x, 276 pp., 78 pl. Meinert, F. °65. Campodee : en familie af Thysanurernes orden. Naturh. Tidsskr. Rek 3, Bd. 3, pp. 400-440, Tab. 14. Translation: On the Campodew, a Family of Thysanura. Ann. Mag. Nat. Hist., Ser. 3, Vol. 20, pp. 361-378, 3 fig. 1867. Meinert, F. 83. Caput Scolopendre. The Head of the Scolopendra and its Muscular System. 77 pp., 3 tab. Copenhagen. 154 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. ; Metschnikoff, E. ; 74. Embryologie der doppeltfiissigen Myriapoden (Chilognatha). Zeits. f. — wiss. Zool., Bd. 24, pp. 253-283, Taf. 24-27. ‘ Metschnikoff, E. | 75. Embryologisches tiber Geophilus. Zeits. f. wiss. Zool., Bd. 25, pp. 313- 4 322, Taf. 20, 21. 4 Miall, L. C., and Denny, A. H ’°86. The Structure and Life History of the Cockroach. 224 pp., 125 fig. London. = Mutr, J. ‘ °77. Uber die Mundtheile der Orthoptera. 30 pp., 8 Taf. Prag. Nassonow, N. 5 87. The Morphology of Insects of Primitive Organization. Studies Lab, Zool. Mus. Moscow Univ. pp. 15-86, 2 pl., 68 fig. [ Russian. ] Nicolet, H. { °42. Recherches pour servir & l’histoire des Podurelles. Nouv. Mém. Soe. — Helv. Sci. Nat. Vol. 6, 88 pp., 9 pl. Neuchatel. : Olfers, E. de. ; 4 62. Annotationes ad Anatomiam Podurarum. Diss. inaug. Berolini. 36 pp. — 4 tab. Oudemans, J. T. ’°87. Bijdrage totde Kennis der Thysanura und Collembola. 104 pp., 3 pl., Amsterdam. Oudemans, J. T. 88. Beitrage zur Kenntniss der Thysanura und Collembola. Bijdr. Dier- — kunde, Zool. Genootsch., — Natura Artis Magistra — Amsterdam. Afley. 16, pp. 147-226, 3 Taf. [Translation of ’87, with brief additions. ] Oulganine [W. N_]. °75. Sur le développement des Podurelles. Arch. Zool. expér., T. 4, pp. xxix-xl. [Abstr. by M. de Korotneff.] Oulianine [W. N.]. °76. Développement des Podurelles. Arch. Zool. expér., T. 5, pp. XVii--XiX. [Résumé by the author. ] Packard, A. S. °71. Embryological Studies on Diplax, Perithemis, and the Thysanurous Genus Isotoma. Mem. Peabody Acad. Sci., No. 2, 21 pp., 3 pl. 6 fig. [Packard, A. S|]. 83a. The Systematic Position of the Orthoptera in Relation to other Orders of Insects. Third Report U. 8. Ent. Comm., ete. pp. 286-345, [7-12], Pl. 23-64. Washington. FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 155 Packard, A. S. *83>. On the Morphology of the Myriopoda. Proc. Am. Phil. Soc., Vol. 21, pp. 197-209, 3 fig. Packard, A. S. *98. A Text-book of Entomology. 729 pp., 654 fig. New York. Palmén, J. A. 77. Zur Morphologie des Tracheensystems. 149 pp., 2 Taf. Helsingfors. Patten, W. *84. The Development of Phryganids, with a Preliminary Note on the De- velopment of Blatta germanica. Quart. Jour. Micr. Sci., Vol. 24, pp. 549- 602, Pl. 36 A-56 C. Rath, O. vom. *86. Beitrage zur Kenntniss der Chilognathen. Inaug.-Diss.Bonn, 38 pp., 3 Taf. Reichenbach, H. *86. Studien zur Entwicklungsgeschichte des Flusskrebses. Abh. Senckenb. Naturf. Gesell., Bd. 14, Heft 1, 137 pp., 14 Taf. Ryder, J. A. *86. The Development of Anurida maritima Guérin. Amer. Nat., Vol. 20, : pp- 299-302, Pl. 15. Savigny, we. 716. Mémoires sur les animaux sans vertebres. Pt. 1, v, 117 pp., 12 pl. Paris. Schaum, H. 61. Die Bedeutung der Paraglossen. Berl. Ent. Zeits., Jhg. 5, pp. 81-91. Sedgwick, A. *88. A Monograph of the Development of Peripatus Capensis. Studies Morph. Lab. Univ. Cambridge, Vol. 4, pp. 1-146, Pl. 1-13. Sharp, D. *95. Insecta. Camb. Nat. Hist., Vol. 5, pp. 81-584, Fig. 47-371. London. Stummer-Traunfels, R. v. 91. Vergleichende Untersuchungen iiber die Mundwerkzeuge der Thysanu- ren und Collembolen. Sitzb. Akad. Wiss. Wien, math.-naturw. Cl., Bd. 100, Abth. 1, Heft 4, pp. 216-235, 2 Taf. Taschenberg, E. L. 79. Praktische Insecten-Kunde, etc. Theil 1, vi, 233 pp. Bremen. Tullberg, T. 72. Sveriges Podurider. Kongl. Svensk. Vetens. Akad., Handlingar. Stock- holm. Bd. 10, No. 10, 70 pp., 12 Taf. Uzel, H. 974. Vor aufige Mittheilung iiber die Entwicklung der Thysanuren. Zool. Anz., Bd. 20, pp. 125-132. 156 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Uzel, H. 97>. Beitrage zur Entwicklungsgeschichte von Campodea staphylinus Westw. Zool. Anz., Bd. 20, pp. 282-237. Uzel, H- 98. Studien iiber die Entwicklung der apterygoten Insecten. vi, 58 pp., 6 Taf., 5 Fig. Berlin. Vayssiére, A. °82. Recherches sur l’organisation des larves des Ephémérines. Ann. Sci. Nat., Sér. 6, Zool., T. 18, 137 pp., pl. 1-11. Viallanes, H. °87. Etudes histologiques et organologiques sur les centres nerveux et les organes des sens des animaux articulés. Ann. Sci. Nat., Sér. 7, Zool., T. 4, 120 pp., pl. 1-6. Walter, A. °85. Beitrage zur Morphologie der Schmetterlinge. Jena. Zeits. f. Naturw., Bd. 18, pp. 751-807, Taf. 23, 24. Westwood, J. O. °39. An Introduction to the Modern Classification of Insects, etc. Vol. 1, xii, 462 pp. London. Wheeler, W. M. °89. The Embryology of Blatta germanica and Doryphora decemlineata. Jour. of Morph., Vol. 3, pp. 291-386, Pl. 15-21, 16 fig. Wheeler, W. M. °93. A Contribution to Insect Embryology. Jour. of Morph., Vol. 8, pp. 1- 160, Pl. 1-6, 7 fig. Wood-Mason, J. "79. Morphological Notes bearing on the Origin of Tne Trans. Entom. Soe. London, 1879, pp. 145-167, 9 fig. Woodworth, W. McM. °93. A Method for Orienting Small Objects for the Microtome. Bull. Mus. Comp. Zodl., Vol. 25, No. 3, pp. 43-47. Zograff, N. ’'83. Marepiarn Kb no3sHaHiW sMOpionasbnaro paspuria Geophilus ferrugineus a eee ee ae ee ee eee ee ee ee ee a ee . L. K. nu Geophilus proximus L. K. Tpygmu .Ja6. mpm Soozor. Myseb Mocroscraro Yuusepcurera, Toms 2. Bun. 1. Mssberia Uunep. Odmecr. In6ur. Ecrecr., Anrponoa. w Sruorp. Toms — xliii. Bun. 1. [Contributions to the Knowledge of the Embryological Development of a Geophilus ferrugineus L. K. and Geophilus proximus L., K. Studies Lab. — Zool. Mus. Moscow Univ., Vol. 2, Pt. 1. 77 pp., 108 fig. ] = FOLSOM: MOUTH-PARTS OF ANURIDA MARITIMA. 15 ~I EXPLANATION OF PLATES. All figures were drawn with the aid of a camera lucida, from preparations dt. ec’drm. . —. . gn. inf’e. gn. swe. Wdrm. . ae cis. lab. . lbr. ten... Ing. . ln. v. 4 of Anurida maritima Guer. ABBREVIATIONS. Adductor. lot. Abdominal appendages, mob. 1st, 2d, 3d. mo. rug. Premandibular append- md. . ages. ms’drm. Thoracic appendages, mu. . 1st, 2d, 3d. max.) . Antenna. ma . Articulation, or hinge. nl. gn. Base. ocl. Dorsal arm. Onde: Posterior arm. @. Buccal cavity. 0. p’at Ventral cord. or. Pivot. pd. Chitin. pd’. Clypeus. phy. Coelom (body cavity). pig: Head. pli. Constrictor. pli. or Cuticula. plp Dorsal. pr’ceb Teeth. prd. Depressor. pr). Deutocerebrum. prt.! Dilator. prt. ms Duct. ret. Ectoderm. sng. cp’. Galea. sta. Infra-esophageal gan- stmd. glion. stp. Supra-cesophageal gan- sul. glion. sul. n. Hypodermis. su’Ing Intima. Staats Incisive. sut.m Labium. te.g Labrum. iijiis Lacinia. tri’ceb. . Lingua. v. Linea ventralis. yk. Elevator. Membrane. Corrugated membrane. Mandible. Mesoderm. Muscle. First maxilla. Second maxilla. Ganglionic nucleus. Ocellus. Dorsal organ. (Esophagus. Post-antennal organ. Mouth. Foot. Footstalk. Pharynx. Pigment. Fold. Mouth fold. Palpus. Protocerebrum. Proctodeum. Projection. Lateral protrusor. Mesal protrusor. Retractor. Blood corpuscle. Stirrup. Stomodeum. Stipes. Trough. Neural groove. Superlingua. Suture. Median suture. Germ band. Tentorium. Tritocerebrum. Ventral. Yolk. Fotsom. — Development Anurida, PEATE WL. Figs. 1-6 represent views of the left side of eggs (embryos) at Stages 1 to6 = tively, with the outer egg membrane removed. X 160. . PLATE 1. IM-DEVELOPMENT ANURIDA. Spices aan 1O 50 helen OOS IOA 3990 a i > fe feisel, fith. Bestan BN Fotsom. — Development Anurida. PLATE 2. Fig. 7. View of left side of embryo at Stage 7, with the outer egg membrane removed. X 160. Fig. 8. Ventral aspect of germ band at Stage 1. X 150. Fig. 8a. Ventral aspect of a portion of the germ band at Stage 1 more hight magnified. x 480. Fig. 9. Left aspect of cephalic region at Stage 3. X 480. Fig. 10. Right aspect of cephalic region at Stage 8. X 480. Figures 9 and 10 are from the same preparation. 7 PLATE 2. : us : 5 SOR 2 DOGO! 1) Sobers 2 | o 4 appx. DEVELOPMENT ANURIDA. B Meisel, lith. Boston aypithe Fig. Fig. Fig. Fig. Fig Fig. Fig. Fig. Fig. Fig. Fig. Fig. The last three figures are from the same preparation. Foisom. — Development Anurida. 21. 22. . Ventral aspect of cephalic region of germ band at Stage 3. > 480. . Ventral aspect of cephalic region of germ band at Stage 4. X 480. . Transection of germ band at the labial segment in Stage 3. XX 762. . Posterior aspect of left first maxilla at Stage 3. Xx 480. . Posterior aspect of left second maxilla at Stage 3. x 480. . Left aspect of cephalic region at Stage 5. x 480. ~~ Fie PLATE 3. Sagittal section of labrum at Stage 3. X 762. Transection of germ band at the first maxillary segmentin Stage 3. X 762. Transection of germ band at the mandibular segment in Stage 3. XX 762. Left aspect of germ band at Stage 4. X 480. Ventral aspect of cephalic region at Stage 5. > 480. Posterior aspect of left first maxilla at Stage 5. Xx 480. FoLSOM-DEVELOPMENT ANURIDA, = plicr b = —————————— XS Foutsom. — Development Anurida. Fig. Fig. Fig. Fig. Fig. Fig. 2 PLATE 4. . Anterior aspect of a transection of germ band at mandibular segment in Stage 7. The section being thick shows both mandibles and, behind them, the superlingue. x 762. . Left aspect of embryonic head, represented as if transparent, at Stage 7. x 480. . Ventral aspect of lingua and first maxilla at Stage 7. X 480. . Dorsal aspect of head of right maxilla at Stage 8. X 762. . Dorsal aspect of lingua and superlinguz at Stage 7. X 762. . Paramedian section to show the primitive cephalic ganglia at Stage 5. X 762. PLATE 4. URIDA. wr AN iM DEVELOPME ], ith. Basten B Meise i ll en ——— —— — — a i Fotsom. — Development Anurida. PLATE 5. Fig. 29. Ventral aspect of head, represented as if transparent, at Stage 7. Fig. 30. Anterior aspect of mouth-parts at Stage 7. From the same pre as Figures 24 and 29. x 480. Fig. 31. Paramedian section of head at Stage 7. X 480. Fig. 32. Ventral aspect of left maxilla at Stage 7. X 762. Fig. 33. Dorsal aspect of adult head. x 45. Fig. 34. Anterior aspect of mouth at Stage 8. xX 480. Phate 5. ESOM—DEVELOPMENT ANURIDA. B Meisel, lith. Beston. . Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fotsom. — Development Anurida, 9 9 30. 36. vi. 38. vd. 40. 41. PLATE 6. Transverse slightly oblique section of head at Stage 8. x 480. Dorsal aspect of left mandible of adult. > 150. Dorsal aspect of anterior extremity of adult right mandible. X 480. Dorsal aspect of skeletal structure of internal mouth-parts in situ. > 150. Dorsal aspect of head of left maxilla in adult. x 480. Surface view of finished labrum. X 150. Head of adult insect viewed from the left side. Xx 150. FOLSOM- DEVELOPMENT ANURIDA. PLATE. 6. ’ d. OcL. 8, Meisel, lith. Boston Fousom. — Development Anurida. PLATE 7. =a Fig. 42. Dorsal aspect of completed lingua and superlingue. X 480. Fig. 43. Surface view (ventral) of adult labium. X 150. Figs. 44-50. Transections of internal mouth-parts of adult to show en 2 at to each other and to the buccal cavity. , Fig. 44 is the most anterior of the series; Fig. 50 the most posterior. Figs. 44 and 45 are magnified 350 diameters, Figs. 46-50, 480 diameters. F OLSOM- DEVELOPMENT ANURIDA. gaslen. kA L SUE. SUL, a GRE meet Le B. Meisel, lith. Boston Forsom. — Development Anurida. PLATE 8. Fig. 51. Reconstruction of part of the left side of the adult head, taken near the median plane. X 380. ao purubs da bus, Meisel, lith Boston BR 5D. Prare G,, 9 B Gog og oO Ma ©0004? . SAY ames r. “La i) FOLSOM-DEVELOPMENT ANURIDA. - Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE. Vous XXXL. No.0: —e TS ON THE DREDGING OPERATIONS OFF THE WEST COAST OF CENTRAL AMERICA TO THE GALAPAGOS, TO THE WEST COAST OF MEXICO, AND IN THE GULF OF CALIFORNIA, IN CHARGE OF ss XANDER AGASSIZ, CARRIED ON BY THE U. S. FISH COMMIS- SION STEAMER “ALBATROSS,” DURING 1891, LIEUT. COMMANDER / ANNER, U.S. N., COMMANDING. XXVIII. [ION OF TWO NEW LIZARDS OF THE GENUS ANOLIS FROM COCOS AND MALPELO ISLANDS. By LEeonnHarD STEJNEGER. hed by Permission of MARSHALL McDoNALp and GEORGE M. Bowers, U. S. Fish Commissioners. ] WiTH ONE PLATE. CAMBRIDGE, MASS., U.S. A. : Z PRINTED FOR THE MUSEUM. Novemser, 1900. i ot a = as 2 EOE = -_ a ———— _ —— —E — — = - — a ea — woIsog yiil SpE ath eres ™ sacl io STIONY No. 6.— Report on the Dredging Operations off the West Coast of Central America to the Galapagos, to the West Coast of Mexico, and in the Gulf of California, in charge of Alexander Agassiz, carried on by the U. S. Fish Commission Steamer “ Albatross,’ during 1891, LizuT. ComMANDER Z. L. TANNER, U.S. N., Commanding. XXVIII. Description of two new Lizards of the genus Anolis from Cocos and Malpelo Islands. By LroNHARD STEJNEGER. The two Anoles here described were the only reptiles obtained on the islands of Cocos and Malpelo during the expedition. Each species is peculiar to the island upon which it is found. Of the two, the one from Malpelo seems to be most highly specialized, there being no nearly related species on the mainland with which I am familiar, while the species from Cocos Island belongs to a group which has a number of representatives in Central America. The two species are only distantly interrelated, inasmuch as they belong to widely separated sections of the genus. It is quite possible that a more thorough search on Cocos Island might reveal additional reptiles. In fact, Mr. Townsend informs me that he saw a snake there which escaped. Anolis agassizi,! sp. nov. Diagnosis. — Tail cylindrical, without crest or keel; dorsal scales keeled, subequal to those on the flanks, slightly smaller than the ventrals, and separated from each other by one or more rows of minnte granules ; ventral scales keeled ; digital expansions very large; about thirty-six trans- verse lamellee under ii and iii phalanges of fourth toe: occipital scale about as large as ear-opening; scales of supraorbital semicircles very much enlarged (forming high, tuberculated crests in the adults), and separated by one row of small scales; occipital separated from supraorbital semicircles by one or two series of scales; supraocular scales rough or rugose, sometimes irregularly keeled ; canthus rostralis sharp; mental shield single, with a deep suleus posteriorly, very large ; tibia nearly equalling the head in length, and at 1 Named in honor of Professor Alexander Agassiz. VOL. XXXVI. — No. 6. 162 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. least as long as distance from mouth to ear-opening ; scales on under side of tail twice as large as those on the upper side, all keeled ; scales on dorsal surface of hands and feet multi-carinate. Adults with a high longitudinal cervico- nuchal flap ; males with enlarged post-anal scales. Habitat. — Malpelo Island, Pacific Ocean, off Columbia, South America. Type. —U. S. National Museum, No 22101; March 5, 1891; collector, Chas. H. Townsend. Description. — @ ad. U.S. Nat. Mus. No. 22101. Head once and two thirds as long as broad, slightly longer than tibia; frontal and occipital regions deeply concave ; supraorbital ridges high, bony, sareoet die the occipital hollow, and nearly joining behind it at the beginning of the cervico-nuchal fold; anteriorly — they divide and continue mesially as frontal ridges which converge on the snout, meeting some distance behind the level of the nostrils, while externally — they join the supraciliary ridge, and in company the latter extend to under the — nostrils as a strong canthus rostralis, thus forming a deep valley on each side between the canthus and the frontal ridge; there is also a post-superciliary ridge extending to above the ear-opening, and with a valley between it and the occipital ridge; scales of supraorbital semicircles very much enlarged, forming high tuberculated bony crests, separated by a single series of very small scales; scales forming frontal ridges and valleys rather large, irregu- larly hexagonal, concave or convex according to situation; scales on snout smaller, more irregular, elongate, four in contact with rostral; about seven larger supraocular scales, keeled or tuberculated, separated by one row of granules from semicirculars; superciliary edge with two very elongated scales anteriorly, granular posteriorly; occipital scale slightly larger than ear- opening, separated from supraorbital semicircle by one row of scales; three canthal scales; loreal region with two deep hollows, the posterior one largest ; loreal rows four; a series of large suboculars, of which the one below the posterior angle of the eye descends to the edge of the lip; rostral very wide and very low, four times as wide as high, nearly rectangular ; six to seven low supralabials in front of the subocular edging the lip, decreasing in height poste- riorly; ear-opening rather small, oval, vertically oblique; nape and neck with a high, flexible dermal crest or flap on the middle line, almost co-extensive with the poorly developed dewlap underneath; several dermal folds and wrinkles on sides of neck ; mental shield large, with a deep suleus behind ; gular scales small, feebly keeled ; body feebly compressed ; dorsal scales slightly larger than those on flanks, a few series along the median line de- cidedly, though not abruptly larger, all more or less distinctly keeled and sur- rounded by one or more minute granules; ventral scales slightly larger than dorsals, rhomboidal, imbricate, keeled, about five to six in the distance between nostrils ; scales on anterior surfaces of limbs larger than ventrals, keeled, those on dorsal surface of hands and feet multi-carinate; adpressed hind limb reaches halfway between eye and nostral; digital expansions very large, thirty-six transverse lamellae under ii and iii phalanges of fourth toe; tail less than twice the length of head and body, cylindrical, without crest or keel; scales on tail larger than ventrals, straight, in transverse rows, but with q 5 Zz STEJNEGER: LIZARDS FROM COCOS AND MALPELO ISLANDS. 163 scarcely an indication of verticels, those on the lower surface nearly twice as large as those above ; a pair of enlarged post-anal scales. Color of live speci- men (according to the sketch of Mr. Magnus Westergrer, the artist of the expedition): top and sides of head and neck uniform sooty black gradually merging into the ground color of the upper surface of body, which is “ Van- dyke ”’ brown, sprinkled with minute dots of an ochraceous buff; upper sur- face of limbs as well as alternate cross-bands on tail similarly colored ; the hands and feet as well as the intervals between the crossbands pale “ Nile” blue; end of snout, lips, and entire under side similarly bluish white. In alcohol the ground color is more blackish and the dots less yellowish. DIMENSIONS. Bi tislen eter. ek bree ek oy ies) oe ee POMLOTear-OPCMING: sy aa. | ss py vn 6200S Snout to vent et Thon ca fo Ben) See h Exel Oe MEIMELOMIGGCNET =. a.) fs ce ww ee fe s. EOE PPM et ee Ps Se BeOS Sereno: Sys fe MOSS S90 © Deere. ney 6 tek) Ce 2 5 Sees BEF Variation. — A large full-grown female (No. 22103) differs from the male described above only in the absence of enlarged post-anal scales. Two some- what younger specimens (female, No. 22104, male, No. 22105) differ from the fully adult specimens chiefly in the lesser elevation of the cephalic crests and the total absence of the cervico-nuchal flap; the color of the back, which seems to be identical with that of the adults, extends also over the upper ‘surface of neck and head. Remarks. — Mr. Charles H. Townsend, who collected these specimens in Malpelo, informs me that they were running over the rocks near the water. The island was too steep to afford a landing, but the lizards were shot off or whisked off the face of the cliffs, thus falling into the water, whence they were secured by the collector. Anolis townsendi,! sp. nov. Diagnosis. — Tail subcylindrical ; dorsal scales but indistinctly larger than those on the flanks, those on the vertebral region keeled ; gular and ventral scales keeled; digital expansion strongly developed ; occipital scale larger than ear-opening, separated from supraorbital semicircles by two or three scales, the semicircles separated by a similar number of scales ; scales on upper surface of snout as well as enlarged supraoculars keeled; anterior half of superciliary ridge with three very long and narrow, strongly keeled scales placed obliquely ; no markedly enlarged series of scales below infralabials ; tibia measuring more than two thirds the length of head, slightly shorter than distance between end of snout and ear-opening; the adpressed hind limb reaches beyond the eye ; tail more than once and a half as long as head and body. 1 Named in honor of Mr. Charles H. Townsend. 164 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Habitat. — Cocos Island, Pacific Ocean, off Costa Rica, Central America. Type.— U. S. National Museum, No. 22106; Feb. 28, 1891; collector, Charles H. Townsend. Description. — @ ad. U. 8. Nat. Mus. No. 22106. Head twice as long as broad, longer than tibia; forehead slightly concave; frontal ridges nearly obsolete; upper head scales small, keeled ; scales of supraorbital semicircles moderately enlarged, separated by three scales; enlarged supraorbitals numer- ous, elongated, sharply keeled, in contact with semicirculars ; occipital shield elongate oblong, somewhat larger than ear-opening, separated from semicireu- lars by two scales; canthus rostralis very distinct, of six scales, the anterior ones small, the posterior two very long and narrow continued backwards in~ line with three superciliaries, which are also unusually long, narrow, and keeled ; posterior half of superciliary ridge granular ; a series of enlarged sub- oculars, keeled, not reaching lip ; loreal rows, about six, keeled ; eight supra- labials to below centre of eye, rugose ; ear-opening moderate, vertically oval; dewlap moderate with a thickened edge of densely set short thick scales, those on sides of appendage distant and very elongate; gular scales small, long, and narrow ; dorsal scales much smaller than ventrals, and indistinctly larger than — those on the flanks, and gradually but slightly increasing in size toward the vertebral line, where a few rows are distinctly keeled; no dorsal or cervical fold or crest; ventral scales larger, imbricate, keeled, like all the scales of the underside ; scales on anterior surfaces of limbs somewhat larger than ventrals, keeled ; tail subcylindric, scales about the size of ventrals, keeled, with hardly an indication of verticels; body compressed ; adpressed hind limb reaches beyond eye ; no enlarged post-anal scales. Color above dull brownish gray, irregularly and indistinctly mottled with dusky which shows a tendency to form cross-bars on the tail; limbs more brownish with lighter roundish spots ; lores, temples, and sides of neck anteriorly with irregular white mark- ings ; a very distinct white, black-edged lateral band from sides of neck over the shoulder to groin; underside whitish ; throat with indistinct, brownish mottlings. ae : DIMENSIONS. Total length 5 4. <- 101 ly . ae ee Snout'to earopening . . 5. 5 4 & - © deem Srout to vent... °c et lk. secs) fo kn Tail'from vent: °° +... .-2 2)" ae oe Fore limb? “28° ete ee Se hee Hind limb: “). = eR ee eo Tibia Ss BAAS, Variation. — A slightly smaller female (U. S. Nat. Mus. No. 22107) differs: chiefly in the absence of a dewlap and in Saleeiaune the white lateral band is — present, but it is not edged with blackish, and there is in addition a narrow white vertebral band from occiput to root of tail. Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE. Vou. XXXVI. No. 7. THE OTOCYST OF DECAPOD CRUSTACEA: ITS STRUCTURE, DEVELOPMENT, AND FUNCTIONS. By C. W. PRENTISS. ‘ WitH TEN PLATES. CAMBRIDGE, MASS., U.S.A.: PRINTED FOR THE MUSEUM. Jury, 1901. ht ee | Pele tat ew CY , ee De =) No. 7.— Contributions from the Zodlogical Laboratory of the Museum of Comparative Zodlogy at Harvard College, under the direction of E. L. Mark, No. 128. THE OTOCYST OF DECAPOD CRUSTACEA: ITS STRUCTURE, DEVELOPMENT, AND FUNCTIONS. By C. W. Prentiss. TABLE OF CONTENTS. IIPHGIMEOM a cps hs) ss 6 es) ws, et te 168 SememmeeMOrpuOlogy . « » « « » © » » « « « « «» 169 BeTINDGEICHIOSUEVOY . « © « «© « « » 6 « «+ « 169 BUPSEMEDEIONIS EL Ta's cc (s felns) 6 3's ee ee, LIT Spe WALT eo oo 0's) 1) 0) yo. ie ens (8 Ne eid orn souk d| PENS eee Sit ta eis takes ss) cs he re st LES sa siructure and Development . ..... =. . 180 PAGE PAGE PAGE PAGE I. Palemonetes. . . 180) II. Crangon 196| III. Cambarus 200|IV. Carcinus 204 1. Structure of the Otocyst . . 180 Pate Maiot oh LOG op) ou eek on 200, eee ts 205 Peeicees. + « 180 + Oe) omer le ees irene 200 pee kemet (20D 6. Sensory Cushion 181 petemtcn op LO sal a het diel ie OD wie is 206 c. Structure of Haws. .- . 182 SUPCMMMicePeLOP | vsy ven aed votes) CZOL Hers wee d, Formation of as. «lw «(182 5 op eo Be ais 5 eg fa Be é. Otoliths . . . .184 aes a OS eg A a IWR 2. Innervation of the Otocyst . . 185 5. ee SOT SORE a Saas A GUS oe Zbl a.Number of cerns, one opie ts each Nerve Ele- ments to a Single Bristle 186 tee oa. 198 Fi adios oem ie-: 5 ee ice Bab 6, Peripheral Ter- Mminations . 188 pees tse”. L599 time tte 0a saat oaem ole ¢. Central Termi- arena) 6. Cg Cw Cw CDC ww ww «208 A tp 2s: d. Histology of the Nerve Ele- TS C) "i (a1: | A a ate eB 8. Development of the Otocyst. (In Homarus americanus.) 193 a. Ist larval Stage 193 ee 194 3d“ = 194 d.4th “© « 495 a VOL. XXXVI. — No. 7 a er Nolet) 204 2 oy ot tet ee hee, (RSE Zoea, 2) ane Sinet t 2Oly, ee A eat! ae £ « ald: ° «© « . « Megalops . 214 7 « « « e Youngerab 214 co a 168 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. PAGE C. Theoretical Considerations. . . 215 2 Comparison of the Otocyst with the Wertclinde Ear 215 : 2. Neuron: Theory 7... dork. ic, xr ee RE < Part II; — Physiology js. «ee % & esas ie) Se PAGE PAGE A. Historical'Survey.- < . = 2°. 215 c. Both Eyes blinded and B. Experiments and Observations . 223 both Otocysts removed . 230 I. The Otocyst as an epee Or- d. One Eye blinded and both ° PANT ye Mente AG 6 Bee Otocysts removed . . . 231 Methods <2 i. .12 223 e. Both Eyes blinded and one 1. Responses of Paleaonetes is Otocyst removed . . . . a Vibrations transmitted to 2. Removal of Sense Organs Waters os) 2 4G vo) P28) and its Effect on the Com- @ a. Normal Gauaniens eet per aoe pensation Movements of the 6. Poisoned with Strychnine 224 IBses: sees -, (eee 932 c. Both Otocysts removed . 224 a. Normal Animals o 2 0 « aoe d. Removal of Antenne and 6. Both Eyes blinded . . . 2382 both Antennules . . . 224 c. Both Otocysts removed . 233 e. Meaning of these Exper d. Both Eyes blinded and TEN e yh ee ic 225 both Otocysts removed . 233 2. Responses of Gulaninins oe 3. Equilibration of Animals nor- ¥ lator); 4G) 2. 226 mally without Otocysts. . 2384 a. To Vibrations: trabeniitted 4. The Effect of the Develop- to Water .. « 226 ment of the Otocyst on the 6. To Atmospheric Boat - 297 Equilibration of Lobster ‘a II. The Otocyst as an Organ of Larve .. 234 Equilibration . . 228 5. The Function of ‘the Otoliths 237 1. The Removal of Sense De 6. The Function of the Hairs of : gans and its Effect on Equil- the Otocyst ~. 5 2 «naam ibration . . « «+ - . 230 | Summary . 2 « «3 o) ee a. Eyes blinded . . . . . 230 | Bibliography . . » ©. 92) 00) ssp b. Both Otocysts removed . 230 | Explanation of Plates... .. . 251 INTRODUCTION. Since the appearance of the admirable paper by Hensen (’63) on the auditory organs of decapods, a period of thirty-seven years has elapsed, a period rich in zodlogical discoveries and improvement in general tech- nique. The great advances made in comparative neurology by means of modern methods have reopened to investigators fields for research hitherto considered exhausted. The zodlogist of the present time is thus enabled to reap a second crop on ground already carefully gleaned, and to harvest results as important as those originally obtained. The physiological work of Hensen’s paper has been continued in re- cent years by various investigators. But aside from the paper by Bethe (’95) on the otocysts! of the schizopod Mysis, little work has been done on the morphology of the decapod ear since 1863. 1 Throughout this paper the terms otocyst, statocyst, ear, and auditory sac _ will be used synonymously to designate the auditory organ, so-called, of Crustacea | PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 169 To throw more light on our knowledge of the vertebrate ear, com- parative study of the (perhaps) analogous organ found among inverte- brates may be of great practical value. For by such comparative study zoélogists have been enabled to solve many perplexing questions which might otherwise have proved too difficult for solution. The present study was undertaken with this practical bearing of the subject in mind, and with the hope that by the aid of modern neurologi- cal technique it would be possible to go deeper into many undecided questions than Hensen could. The work is necessarily twofold in its scope, owing to the inseparable nature of the morphology and physiology of the auditory organ. We have, first, to obtain more accurate knowledge concerning the structure, innervation, and development of the decapod otocyst. In doing this especial attention must be given to the innervation, which must be com- pared with that of other sense organs in decapods. And, secondly, we must determine from evidence obtained by others in the past, and from additional physiological experiment, whether we are justified in ascribing a true auditory function to this much discussed apparatus. PART I.— MORPHOLOGY. A. HISTORICAL SURVEY. Although the literature up to Hensen’s time is well summarized by him, yet it may be worth the while to take a glance at what has been done, touching upon only the more important works, however, as a fairly complete list of authors is appended in the Bibliography. The earliest notice of an ear in Crustacea is that of Minasi, a Domini- can monk, who in 1775 attributed the sense of hearing to Pagurus, the hermit crab, and described as the auditory apparatus what is now known as the green gland or excretory organ of decapods. The organ supposed to subserve the function of hearing was thus from the very first mis- placed, and its identity was in doubt even up to the time of Hackel (57) and Leydig (’57), who were the first to rectify the erroneous ideas which existed in regard to the functions of the green gland and the otocyst. The true sacs were, however, discovered and described as early as 1811 by Rosenthal (11). He mentions the cavity, its opening, and nerve; but it was left for Treviranus (’02-’22, Bd. 6, pp. 308-310) to discover the sand, or otoliths, present in the otic chamber. 170 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. The first good description of the organ, accompanied by figures, was given by the Englishman Farre (’43), who carefully dissected the otocysts of the crayfish (Astacus fluviatalis), the European lobster (Astacus marinus), the hermit crab (Pagurus), and the rock lobster (Palinurus quadricornis). The organs were found by Farre to be situated in the basal segment — of the inner antennz (antennules), the thin dorsal membrane of which } in A. marinus he compared to the fenestra ovalis of the vertebrate ear. — The openings of the sacs were always found to be large enough to admit the otoliths, which rest upon auditory bristles. The otoliths were, he — maintained, merely grains of sand. The auditory bristles were briehay described, and their semi-circular arrangement noted; a nerve was — traced from the brain to the ventral surface of the otocyst, where it formed a plexus. In Farre’s opinion separate fibres probably supplied | the bases of the different hairs. While the otocysts of the lobster, crayfish, and hermit crab were of relatively large size, nearly filling the basal segment of the antennule, their openings were very small and well guarded by a “chevaux de frise” of bristles. In Palinurus the organ was apparently degenerate ; the sac small, shallow, with very large opening, and the auditory hairs sparse and irregularly arranged. The otoliths were of large size and few in number. The whole apparatus” was held by Farre to be a delicately modified tactile organ, and he doubted if a true auditory function could be ascribed to it. During the next twenty-five years otocysts were discovered and ex- amined in various decapods by Souleyet (’43), Von Siebold (’44, ’48), Leuckart (’53, ’59), Frey und Leuckart (’47), Huxley (’51), Leydig (755, ’57, 60), Bate (’55, 58), Hensen (63), Sars (’67), and Lemoine (’68). Leuckart und Frey (’47) briefly described the sacs which they found in the endopod of the last abdominal appendages of Mysis, mentioning the otolith and auditory hairs. Leuckart (’53) made a comparative study of the otocysts in many crustacean forms. He divided them into two groups : — Those having (1) closed sacs with one otolith, and (2) open sacs with many otoliths. Leuckart’s general descriptions agree with those of Farre. Kroyer (’59) devotes a few pages of his monograph on Sergestes to a comparative account of this organ in different Crustacea. He follows Leuckart’s method of grouping. To the first type (closed sacs, and one otolith) belong such forms as Lucifer, Sergestes, Mysis, and Phyllosoma, In the second group (open sacs and many otoliths) are placed Homarus, Astacus, and Palinurus. In the opinion of Kroyer, Farre erred in con- PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. eL sidering the otoliths simply particles of sand; for sometimes the sacs are closed, and again the openings are often too small to admit the passage of the otoliths from the exterior. They must be, then, deposits of calcium carbonate secreted by the animals themselves. Hensen’s (’63) account of the otocyst is far more complete than any other, and a fairly extensive review of his paper is necessary for the sake of later comparisons. He worked mostly with freshly collected animals, although some twenty-four species were studied from alcoholic material. His principal methods were dissection and maceration, some few crude sections, however, being made. The paper is divided into an anatomical and a physiological part. The latter portion will be re- viewed, along with other papers of a similar nature, in Part II of this paper. The elementary parts of the typical auditory organ are described by Hensen (63, p. 326) thus: “Der Gehérapparat der héheren Krebse besteht nun, kurz gesagt, darin, dass, von der Endganglie eines Nerven ein feiner Faden in ein Chitinhaar hineintritt, und an einen eigen- thiimlich gebildeten Theil der Haarwand sich festsetzt. Diese Haar- wand ist so Jocker mit der Schalenhaut verbunden, dass sie bei entsprechenden Toénen recht bedeutende Schwingungen vollfiihren kann und vollfihrt. Das Haar selbst geht zuweilen noch in oder zwischen Steine hinein.” Crustacea he divides into four classes according to the condition of otocyst and otoliths :— 1. Sacs closed, with one otolith : example, Mysis. 2. Saes closed, without an otolith: all Brachiura. 3. Sacs open, many otoliths : Astacus, Paleemon. 4, No sac nor otoliths, but free auditory hairs. Otoliths. In confirmation of Farre it was found that the otoliths of decapods having open ear sacs were mainly composed of grains of sand. This was proved by chemical tests, and by keeping newly moulted animals (Palemon) in filtered water to which uric acid crystals had been added. Examination of the otocysts some time after moulting showed the presence of these crystals in the sac. In larger forms, such as the lobster and crayfish, the sand particles are spread over the whole basal surface of the ear sac. In shrimps and prawns they are more closely aggregated. The single otolith found in Mysis flexuosus is described at length, but as this account has been corrected by Bethe (795), it will be referred to later in connection with Bethe’s work. The Otocyst (Hérblase of Hensen) is described in general as a round- a 172 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. ish or ovoid cavity, lined with chitin; the opening, if any, is always dorsal, and varies greatly in size. It is found in the basal segment of the first antennz of all decapods, and in the endopod of the sixth or last abdominal appendage of the schizopods. The sac is closed in the Bra- chiura and Schizopoda, but open in most Macrura. The otocysts of Crangon, Palemon, Hippolyte, Mysis, and Carcinus mzenas are described in more detail, but no good figures or sections are given. Auditory Hairs or Bristles. Hensen gives the first and only good description of these. They differ from common tactile hairs in that the. hair shaft is not directly connected with the wall of the sac, but a thin chitinous membrane intervenes, forming a small hollow sphere. It is this ‘‘ spherical membrane” which allows the great freedom of move- ment necessary for the shaft in its response to sound vibrations. A peculiar process, the “lingula,” projects from the inner wall of the base of the shaft into the spherical membrane, and to this the nerve fibre is attached. The hair shaft is generally plumed, as in tactile hairs, with delicate chitinous filaments. In A. marinus the hairs are plumed and are nearly one millimetre in length. They are here very numerous, 468 having been counted in one case, and are arranged on the floor of the otocyst in four parallel semi- circular rows. ; A. fluviatalis has a much smaller number of hairs, but the same general arrangement ; Crangon, a row of only seven or eight ; these are more attenuate than in either of the above forms, but are 0.75 mm. in length. Palemon antennarius has about 40 hairs, arranged in a half-oval or horseshoe shape, the break in the oval being posterior. The hairs them- selves are peculiar in having their shafts bent at a sharp angle. The portion of the shaft above the bend is much longer and more attenuate than the basal part, and. is also heavily plumed. These plumed ends project toward the centre of the horseshoe, and intertwine. Their length is about 100 and their greatest diameter 3.8 u. The hairs of Hippolyte and Mysis strongly resemble those of Palemon, but they are embedded in the single otolith and are therefore unplumed. Carcinus menas has about three hundred auditory hairs. They are grouped into three classes: — 1. Hook hairs (Hakenhaare): the shaft hooked and with a plumed tip, about thirty in number, 50 u long, similar to the otolith hairs of Macrura. 2. Thread hairs (Fadenhaare) : long, filamentous, plumed at very tip, a single row of about 46, each 338 a long, 3 in diameter. 3. Tuft hairs (Gruppenhaare): short, blunt, and unplumed, about 200 in number, occurring in a single large group. PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 17. Hensen also found on the appendages of some decapods free hairs which closely resembled auditory bristles, and are described as such by him. Crangon especially, which has few hairs in the otocyst, is supplied with many of these so-called “free auditory hairs.” They are also numerous in Mysis and Palzmon. Innervation of the Otocyst. In Paleemon Hensen traced the nerve of the first antenna from the brain. A large branch of this nerve runs to the ventral side of the otocyst, where the fibres separate, each enlarging into a ganglionic cell and then proceeding to the base of a hair. Each of these terminal fibres (‘‘ Chord ” according to Hensen) then enters the pore beneath a hair, passes through the spherical membrane to the lingula, or process from the base of the hair shaft, and makes itself fast to this. In his own words (Hensen, ’63, p. 368): ‘ Dieser eigenthiim- liche Faden, den wir als Chorda bezeichnen, lauft eine kiirzere oder langere Strecke weit bis zu einem Horbaare hinfort, und geht durch die Mitte des Porenkanals und der Haarkugel bis zur Lingula hin, an die er sich festsetzt.” Essentially the same conditions were found by Hensen in Carcinus mznas and in Mysis. He also found nerve fibres supplying the tactile bristles which are present on all parts of the decapod body. Formation of New Hairs (Haarwechsel). New hairs are not formed inside the old, but beneath the chitinous wall ; and instead of developing from a single matrix cell, as was supposed, Hensen found that each was the product of a great number of cells. A new layer of chitin is formed beneath the old, and wnder this new layer, but continuous with it, the new hairs are formed as double-walled (i. e. invaginated) tubes. The new chitin wall is compared to the hand of a glove. If the fingers of the glove be turned partially outside in, so as to leave only their tips projecting, the condition would represent that of the hair tubes just before the moulting of the old shell. The tips of the newly formed hairs become attached to the shaft of the old hair, into which they project some distance, and as the latter are detached at ecdysis, the new hairs are pulled out. Nerve fibres were found running into the very tips of the new hairs. MHensen’s theory is, that at moulting, the old nerve fibre, becoming more highly refractive and resembling chitin, is, upon the detachment of the old hair, drawn out through the apex of the new one, and that before this event a new fibre is formed. This theory, however, is not easily reconcilable with his statement that the nerve fibres attach themselves to the lingula at the base of the hair shaft. The remainder of this part of his paper is devoted to brief descriptions of the otocyst as found in some twenty-four different species of Crustacea. To this is added a table, embracing all the forms which have been stud- ied, giving the names of the different investigators, and the conditions, as to number and size, of both the auditory hairs proper (Otolithen- haare) and the “free auditory hairs ” found on the antennz and abdomi- ~ nal appendages. Lemoine (’68) compares the otocyst of the lobster with that of the crayfish. His descriptions are similar to those of Farre (43), but his figures are poor. The thin dorsal wall of the basal segment of the first — antenna, which covers the ear sac, he calls the tympanic membrane of ; the lobster. The opening of the sac is overlooked, and the otocyst de- scribed as closed. Thus, as the otoliths cannot come from without, Lemoine’s theory is that they are exfoliations from the calcified walls of the sac, — an absurdly impossible assumption, as the thin chitinous walls” of the otocyst are not calcified. In the case of the crayfish, he notes’ nothing new except that there is a membrane at the base of each hair shaft, separating its cavity from that of the spherical enlargement on — which the shaft stands. This membrane acts as an ear drum, taking the © place of the large tympanic membrane described for the lobster. Garbini (80) discusses very briefly and incompletely the sense organs of Paleemonetes varians. The figures of the otocyst are extremely crude considering the date of the work, and simply confirm the conditions found — by Hensen in Palemon. ; Vom Rath (’87, ’88, 91, 94) does not make the sharp distinction between auditory and tactile hairs that Hensen does, holding that the two kinds grade insensibly into each other, the auditory hairs being simply slightly modified tactile organs. All sensory bristles of Crustacea can be divided into two chief groups : — . (1) Tactile or auditory hairs, with long, plumed shaft, the base of which is attached to the body wall by a delicate membrane of chitin, often spherical in form. Differentiation is thus towards freedom of movement in response to tactile or vibratile stimuli ; (2) taste or olfac- tory hairs, having a short blunt shaft, thick-walled at the base, but with either a small pore or thin permeable membrane at its distal end, by means of which chemical substances in solution can come into direct contact with the nerve endings. The nervous apparatus of these hairs is the same in both cases for all decapods. The sweeping statement is made, that beneath every sense hair there lies, either in the hypodermis, or removed some distance from it, a group of bipolar ganglion cells. From each of these cells a fibre is given off peripherally, and these, forming @ strand, enter the base of the hair, ending only at its very tip. , 174 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. ‘ PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 175 Claus (’75, 91) agrees with Vom Rath as to the nerve ending, but maintains that there is only one ganglion cell sending its process through a group of matrix cells into the hair. A misunderstanding as to the relations of the ganglion and matrix cells forms the basis of several con- troversial papers. Retzius (90, 92, ’95) concludes in his last paper (’95) that there may be several ganglion cells to a single sensory hair. The number may indeed vary from one to many. He was unable by any method to trace the peripheral nerve fibres further than the base of the hairs. Nerve endings, which he described in his first paper (’90) as extending into the hair shaft, he afterwards (’92) frankly acknowledges to be artifacts. Bethe (’95°), in his admirable little paper on the otocysts of Mysis, clears up by modern methods many points, and corrects some of Hensen’s erroneous descriptions. The sac in Mysis is ellipsoidal, and pointed posteriorly, while from its floor rises a sensory cushion bearing the hairs. This cushion is tilted outwards and ventralwards 45°, the right and left cushions thus being perpendicular to each other. The sac is open, not closed as described by Hensen; the narrow aperture is con- cealed by the overlapping walls of the otocyst. Borne on the sensory hairs is the large otolith, oval as seen from above, kidney-shaped in side view; its greatest diameter 0.3 mm., the other dimensions being 0.25 mm. and 0.15 mm. It is composed of a more or less organic core, about which concentric layers of calcium fluoride are deposited. The tips of the sensory hairs are embedded in this inor- ganic layer, and penetrate to the core of the otolith. The layers of calcium fluoride are probably deposited from the sea water. The sixty sensory hairs are arranged in a single row, so as to form two thirds of a circle, the break in the line being posterior and toward the median plane of the animal. At one end of the curve five hairs are grouped together, and at the other end there is an irregular double row. Though much like the auditory hairs of Paleemon, their tips, em- bedded in the otolith, are unplumed. Only one ganglion cell to a hair was found, sending a distal process into the base of each shaft. A double row of matrix cells lies just beneath the single row of hairs, and could easily be mistaken for ganglion cells. Vom Rath may have made this mistake, thus getting a multiganglion-celled condition for each hair. : The otocyst begins to develop before the appendage is fully formed. An invagination of the dorsal ectoderm takes place, producing a shallow depression ; this enlarges while the opening gradually closes. Certain 176 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. of the hypodermis cells elongate to form the matrix cells which later produce auditory hairs. The latter are formed only after hatching. Herrick (95) mentions the auditory organ only in connection with the development of the lobster. The otocyst becomes prominent at the third larval stage, appearing as a shallow depression bordered with short setee and containing a few grains of sand. The depression gradu- — ally enlarges, forming in the fifth stage a sac, the aperture of which decreases in size with successive moults, until the adult condition is attained. Bethe (95, 97) has traced the auditory fibres of Carcinus menas — centrally to the neuropil of the first antenna, where they end in delicate fibrillations. Some of these fibres may also end in the globulus. From this review of the literature, it is seen that little has been done on the finer anatomy of the otocyst. Hensen’s work, once considered | exhaustive, will not suffice at the present time. The organ of Brachyura has not been touched upon since Hensen’s dissections, while our knowl- edge as to the innervation of the different sensory hairs of Crustacea is left in a very hazy, confused state, since the exact condition of the peripheral endings is not firmly established, Claus, Vom Rath, Retzius, and Bethe each holding different views. The question remains un- settled as to whether the manner of innervation is the same for all the sensory hairs. G. H. Parker (’90) has clearly shown that the optic | nerve in Crustacea is highly differentiated ; but all the other sense organs have, according to Vom Rath, the same manner of innervation, even though they differ in function as much as the so-called auditory and olfactory bristles. All the investigators of the crustacean otocyst, Bethe alone ex- cepted, carried on their work under the impression that they were dealing with an auditory organ. This certainly prejudiced them in drawing conclusions. But for this, Hensen would never have likened the thickened wall of the crab’s otocysts to the malleus of the verte- brate ear, nor made other far-fetched comparisons. A comparative study of the innervation of the otocyst, especially if supplemented by that of the olfactory and tactile bristles and the conditions in embryonic stages, cannot fail to clear up some of these questionable poiuts. PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 177 B. OBSERVATIONS. In the account of the morphology of the otocyst, two types will be taken for description : — (1) Open otocysts containing otoliths (macruran decapods) ; the example will be Palemonetes vulgaris Stimpson. The otocysts of the crayfish Cambarus affinis (Say) Girard, and of the prawn Crangon vulgaris Say, will be described in only sufficient detail to allow of comparison with Palemonetes, and to correct any errors or omissions in the descriptions of other investigators. (2) Closed otocysts without otoliths (brachyuran decapods) ; the sac of the green crab, Carcinus menas Lin., will be taken as the example of this type. For tracing out the development of the macruran otocyst (1), young lobsters were used instead of Palemonetes larve, as it is difficult to obtain a complete series of the latter, and their small size makes them by no means favorable material for studying the embryology of the sac. Young lobsters, however, can be had in abundance during the hatching season, and are of large size; the otocyst is of the same general type as that of Palemonetes. The development of the closed otocyst (2) was traced out in the crab for the sake of comparison with the macruran type of sac. The research represented in this paper was carried on at the sug- gestion of Dr. E. L. Mark, to whom I wish here to express my thanks for his constant kindness, suggestive direction, and able criticism. I am also indebted for valuable supervision and helpful suggestions to Dr. G. H. Parker, who directed my work for one year during the absence of Dr. Mark. 1. Material. Large numbers of Palemonetes were obtained from the Charles River, Cambridge, at low tide. These river animals live well in either salt or fresh water, and may be kept in aquaria without running water for an indefinite period. Being so hardy, and at the same time free swimmers, they are eminently adapted for cztra vitam stains, and for physiological experimentation. Carcinus mznas was abundant in the soft-shelled condition, at Hadley Harbor, Naushon Id., during the months of June and July. The head of Great Harbor, Wood’s Hole, was another good collecting ground. 178 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Many soft-shelled animals were obtained by keeping young crabs in aquaria, and feeding them freely until ecdysis took place. Lobster larve were hatched at the U.S. Fish Commission Station, Wood’s Hole, during June and July. They were reared, but with great difficulty, up to the eighth moult. Fed on minced crab’s liver they throve well; but unfortunately they also fed indiscriminately on each other. Crangon was found in large numbers in the muddy bottom of the Charles River ; crayfish were bought in the New York City markets. 2. Methods. In sectioning, great difficulty was experienced, both on account of the thickness of the chitin, which was often calcified, and because of the siliceous otoliths, so numerous in the sacs of Macrura, and glued by secretions to the hair tips. As the otoliths are insoluble in acids strong enough to completely destroy organic tissues, the only successful remedy was to remove them mechanically. This was best accomplished by washing them out by a stream of water blown into the sac. The apparatus for this consisted of a short piece of small rubber tubing into one end of which was inserted a glass tube drawn out to a fine point. The other end of the tubing being held in the mouth, and the capillary tube inserted into the aperture of the otocyst, a stream of water was driven into the cavity of the sac with considerable force. The larger otoliths having been washed out in this way, fairly good sections could be cut. In the crab, the difficulty in cutting the very thick calcified chitin was obviated by using soft-shelled animals. The chitin is at this stage very thin, uncalcified, and therefore more readily sectioned. Lobster and crayfish antennules were decalcified by placing them in Gilson’s fluid for twenty-four hours, or in Vom Rath’s platinic-osmic fixative for a week or ten days. Of the many fixing reagents used, (1) Vom Rath’s platinic-osmo-picro- acetic mixture, (2) his corrosive-picro-acetic fluid, and (3) corrosive sublimate plus 1 % acetic acid gave the best results and in the order named. The last two were followed by staining in iron haematoxylin, which gave a clear definite stain of sections as thick as 20u. The platinic- chloride fixative of Vom Rath was used for from three to five days, either followed or not by treatment with pyroligneous acid. In Palemonetes and Crangon a fine differentiation of fibre tracts was obtained by using — PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 179 the fixative alone for three to five days, and washing out for at least two weeks in 90 % alcohol. The myelin sheath was intensely blackened, while all other tissues remained a yellowish brown. For tracing nerve fibres, both to peripheral and central endings, intra vitam staining proved of most value. Different methods were employed for obtaining peripheral and central stains. A one per cent solution of methylen blue in normal NaCl was injected into the body in either case. For peripheral endings several injections were made into the abdomi- nal blood space, at intervals of thirty minutes. When the animals showed signs of stupefaction, a final injection was introduced into the pericardial chamber. The amount of solution injected varied from a few drops, in Palemonetes, to five cubic centimetres, in the lobster. In from 15 to 30 minutes after the final injection the animals were usually dead. The part to be studied was then dis- sected out, barely covered with normal salt solution, and examined from time to time under the microscope, until a satisfactory degree of stain- ing had been reached. For central terminations one injection only was made, and this directly into the chamber of the heart, only a few drops of the solu- tion being required. When the blue color was well diffused throughout the tissues (about one hour after injection), the brain was dissected out, or exposed, and examined as before. For fixation of the stain Bethe’s ammonium-molybdate method for invertebrates was used. It was found to be better to leave preparations in xylol for only the shortest possible time, as this reagent diffuses the color. Preparations fixed by this method keep very well for a year or more, but after this they ultimately deteriorate, fibres originally sharp and continuous in outline becoming miere dotted lines, while the surrounding tissues take on a deep yellow hue. When both brain and otocyst were examined together, the peri- pheral cells and fibres stained first, then central fibres, central termina- tions, and ganglion cells of the brain in the order named. Sections 60-120 u in thickness were cut, but by far the greater number of pre- parations were examined in toto. The transparency of the tissues made this possible even with the brains, a millimetre or more in thickness, of large crayfish. To get constantly complete impregnations of both peripheral and cen- tral endings, it is necessary to expose to the atmosphere the part to be studied. The impregnation then takes place sooner, lasts longer, and affects a larger number of elements. The fixation of the color is also much better in this case, because the fluid can penetrate much more 180 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. readily, and on the rapidity of its penetration depends, in a large meas- ure, the success of fixation. Gold-chloride and Golgi preparations were | useful only for supplementing and controlling the results obtained by — methylen blue. Both the rapid and slow processes for silver impregna- _ tions gave fairly good preparations, but by no means as complete or constant results as methylen blue. Ranvier’s gold-chloride method, s a in which formic acid is used for reduction, was very uncertain in a action on nervous tissue, but was quite useful in bringing out fine cell processes in the sensory hairs. ae 3. Structure and Development. I. PALEMONETES VULGARIS STIMPSON. 1. Structure of the Otocyst. - a. Sac. This is situated, as in all decapods except the Myside, in the basal segment of the antennule, nearly filling its cavity. Its out= line as seen from above (Plate 1, Fig. 1) is nearly ovate, being well rounded posteriorly, though suddenly becoming pointed at its anterior end. In individuals of medium size (30 mm. long) its average dimen- sions are 0.66 mm. in length, 0.63 mm. in width, and 0.33 mm. in depth. In longitudinal section (Plate 1, Fig. 4) its outline is somewhat kidney- | shaped, its length being about twice its depth, and its ventral wall projecting into the lumen. ‘Transverse sections through the basal portion of the antennule (Figs. 2, 3) show that the lumen of the otocyst is from one half to two thirds as wide as the antennule at this point. The chitinous wall of the sac, which is extremely thin, is continuous with that of the antennule (Fig. 3). The hypodermal cells form a single layer, except in the sensory region of the sac, where they are elongate and several layers thick. Median to the otocyst passes the antennular nerve, the cut end of which is shown at x. at. J (Plate1, Fig. 2), and directly below it lies the large muscle of the segment. Otoliths occupy the median and posterior portion of the lumen, and nearly conceal from view the sensory hairs (Fig. 3, set. ot.). In para sagittal sections (Fig. 4) is to be noticed the close proximity of the brain (n’ pil. opt.), which is not more than 0.22 mm. posterior to the |~ sac, and projects somewhat into the base of the antennule ; the sensory cushion, or prominence (ers. sns.), bearing the stumps of a few severed | hairs, is also to be seen. The long axis of the otocyst is not coincident with that of the anten- — PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 181 nule (Fig. 1), as its anterior end is more lateral in position than the posterior. The external aperture has the form of a pointed ellipse and penetrates the dorsal wall of the antennule ; it is nearly as long as the sac itself, but does not extend quite as far back as the sac. It was described by some of the early writers as a longitudinal slit, by others as transverse; but, as Hensen points out, it is neither: its direction is oblique, and corresponds to that of the long axis of the otocyst. The opening is completely covered over by a thin fold of chitin (Figs. 1, 3, tet.), which extends forward and laterad to end in a sharp projection or spine. ‘This lid-like fold (tectum) must be lifted or cut away in order to come directly at the opening of the otocyst. Figure 3 shows the position and form of the lid in transverse section, and how closely it fits over the aperture of the otocyst, while its forward projection over the an- terior lip of the slit can be seen in Figure 5 (Plate 2) at tet. As the chitinous lining of the otocyst is of ectodermal origin, like all other chitinous parts, it is cast off at each ecdysis, with all it contains, and a newly secreted sac takes its place. b. The sensory cushion of the otocyst is produced by an elevation of the median and posterior portion of the floor of the sac, which projects into the lumen and gives a somewhat constricted appearance to the cyst in sagittal sections. The surface of the cushion, which is about 0.25 mm. in diameter, is not horizontal, but slants downward from the median side of the sac to its lateral wall at an angle of nearly 45° (Plate 1, Fig. 3). This makes the sensory cushions of the right and left sides per- pendicular to each other, a condition similar to that described for Mysis by Bethe (’95*, p. 556), and of some physiological importance. The sensory hairs are borne on the sensory cushion, and for this reason the prominence has been compared to the criste acustice of vertebrates. The hairs, or bristles (for both names are applied to them), vary from forty-five to fifty-eight in number, and are arranged in a curved horse- Shoe-like row (Plate 1, Fig. 1), the two ends of which are directed obliquely caudad and mediad. Largest at the inner end of the curve, and arranged in a single row, they grow gradually smaller toward the other end of the series, where an irregular double line is formed. Fig. 6 (Plate 2), a transverse section through the posterior ends of the horseshoe shows the base of a single hair on the right or median side, while at the left or lateral end two bristles are seen, the lateral row being double. Directly beneath the hairs we find, instead of the usual layer of VOL. XXXVI.— NO. 7 2 182 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. hypodermal cells, groups of cells with elongated nuclei; these send their processes into the bases of the bristles (Plate 2, Figs. 6,7). They are the matrix cells, which nourish the fair and, as we shall see later, have to do with its formation. The central region beneath the cushion is occupied posteriorly by the ganglion cells of the otocyst nerve (Plate: 2, Fig. 6, cl. gn.), and anteriorly by their peripheral fibres. c. Structure of hairs. The hairs of the otocyst are peculiarly modi- fied. Instead of being straight, as in tactile hairs, the shaft is here bent out of its course about 120°, so that its distal portion makes a sharp angle with the proximal end (Plate 2, Fig. 8). The shaft is very lon in comparison with its diameter, being from 1604 to 200. in length, while only 3 to 6m in diameter at the base. The part of it above the bend becomes extremely attenuate, and is heavily fringed with long deli- cate projections (pinnules), which give it the appearance of a plume. These fine feathery tips, which always project toward the concave side of the horseshoe formed by their bases, are crisscrossed and tangled to- gether in such a way as to form a wickerlike mesh, on which the majority of the otoliths rest (Plate 1, Fig. 3). The hairs are not attached firmly or immovably to the wall of the sensory cushion, but an exceedingly thin-walled chitinous bulb intervenes between the shaft and the wall of the sac. This, the spherical membrane of Hensen, is from 6 to 12 p i diameter, and allows the shaft, itself comparatively rigid, to sway freely on its base, as if articulated there (Plate 2, Fig. 8, mb. sph.). d. The formation of hairs has already been described by Hensen ('63, p. 374) in some detail. The conditions just before ecdysis were figured, but the earlier stages were not given; so a few supplementary facts may be added here. Braun (’75) verified Hensen’s account Haarwechsel in the bristles of Astacus, and himself discovered som: new details. ‘ As before stated, each sensory hair is produced by a number o matrix cells, which send their processes into the shaft. In newly forme: hairs, these protoplasmic processes extend to the very tip of the h i cavity (Plate 2, Fig. 7). In preparation for the next moult they a withdrawn nearly to the base of the hair, leaving the greater part of th hair cavity empty (Plate 2, Fig. 9). At the same time the matrix cells from which these processes are given off sink deeper into the tissue, below . the level of the hypodermis, and with other chitinogenous cells originat- ing in the hypodermis, arrange themselves about the nerve fibre of the PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 183 old bristle for the purpose of forming the new hair (Fig. 9, cl. ma.). This aggregation of cells is similar to the papilla described by Braun, but they are by no means as regular in outline and arrangement as those figured by him. In Palemonetes this condition of the matrix cells exists for several weeks before ecdysis takes place, the new hairs being formed during this period. In adult lobsters and crayfish the time is probably much longer, whereas in larve it lasts but a few days. The chitin of the new hair shaft is secreted pari passu with that of the test, so that the two are continuous, but the hew hair is beneath the shell, in the region where the matrix cells have formed the papilla. It is secreted as a double tube, the distal end of the inner part of which projects, as the tip of the new hair, into the base of the old one. Figure 10 (Plate 3) shows the condition of affairs just before ecdysis in the endopod of the third abdominal appendage ; cta. being the old test, cta’. the new one formed beneath it. Three newly formed hairs are seen as double tubes located deep in the appendage. The walls of the two tubes are continuous with each other at their lower or proximal ends, and the tip of the inner tube projects distally into the shaft of the old hair. This inner tube, the tip of the new hair, must be secreted by the delicate processes from the matrix cells which still extend up into the old hair during the period of hair formation. The outer tube, though continuous at its lower end with the inner, is secreted by two parallel rows of matrix cells, very similar to the chitinogenous cells of the hypo- dermis, and probably derived from them. Hensen (’63, p. 375) has well described this condition of the new hairs as resembling that of the finger of a glove turned partially inside out, the tips alone projecting. The tip of the new hair is embedded in a viscous, homogeneous substance, which is formed between the old and the new tests, either by glandular secre- tion of other cells or by the chitinogenous cells themselves. This substance probably corresponds to the homogeneous non-cellular mem- brane found by Herrick between the shells of the lobster (95, p. 87). When the old test is shed, it adheres to the fine plumes of the new hair tip, and aided by the internal blood pressure (very considerable at the moulting period), draws the recently formed hair out into its functional position, just as one would draw out the invaginated finger of a glove by puiling on its tip. The chitin of the shaft is very soft and pliable at this time, allowing the hairs to be turned right side out with ease; indeed, this may be done artificially. But if by some accident at the time of ecdysis any of the hairs are not at once fully drawn out, the chitin hardens and they are fixed in their abnormal position. 184 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Aside from its general interest, this peculiar method of forming the new hair is very important, as throwing light on the peripheral endings of the nerve fibres in the sensory hairs. By it certain conditions may be explained. At each moult the nerve fibres lose their connection with the old hairs, and come into relation with new ones. How these changes are brought about can best be described in connection with the innervation of the otocyst. e. The Otoliths are borne in a rather compact mass upon the inter- — laced tips of the sensory hairs (Plate 1, Fig. 3, of’/th). They consist of irregular grains of sand mingled with other fine mineral particles and organic detritus. The largest measure from 8 to 12 uw in longest dimen- sion. That the greater part of them are siliceous is shown by their insolubility in strong sulphuric acid, and by the fact that they scratch glass when crushed upon it. They are renewed after each moult, for the freshly formed sac is at first without them. New otoliths are pushed in by means of the chele through the aperture of the sac while its walls are yet so soft and flexible as to admit quite large grains of sand. By watching animals soon after moulting it can be observed that they stir up the sand at the bottom of the aquarium in which they are confined ; as soon as some particles have come to rest upon the dorsal side of the antennule, one or both chelz are raised, and by their tips the grains of sand are pushed back under the protecting lid of the opening into the otocyst. Otocysts from which most of the sand parti- it cles had been carefully removed by forcing a jet of water into the sae were found after a lapse of two days to contain otoliths derived fror a iron filings which had been strewn on the bottom of the aquarium. The otoliths are often entangled in the feathery plumes of the auditory hairs, and are in this case attached to them by an organic substance, which i ; probably secreted by unicellular glands situated beneath the floor of the sac. No multicellular glands, such as are found in the lobster and cray- fish, could be detected beneath the otocyst of Palemonetes. Very minute canals, which are probably the ducts of gland cells, were found running through the chitin wall and some distance into the tissues beneath ; they were very clearly brought out, and their tubular cond? tion proved beyond a doubt, in silver preparations, and in those made with lead formate; but unfortunately their connection with gland cells could not be demonstrated. The functions of the otolith and the part it plays in audition, or equilibration, will be discussed in experimental portion of this paper. | | . . : | . . | | | | PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 185 2. Innervation of the Otocyst. As already noted, the brain, or supra-cesophageal ganglion, is less than a quarter of a millimetre distant from the ear sac. The nerve supplying the hairs of the otocyst is thus comparatively short, and can be traced in a single section from the central to the sensory termination. Figures 4 and 12 (Plates 1, 3) show its general course after leaving the brain. Its sensory ganglion lies directly beneath the posterior end of the sac. The nuclei of the nerve cells of the ganglion are situated about 0.25 mm. back of the hairs which they innervate, grouped irregularly together ; the peripheral fibres of the cells run somewhat parallel to one another, then spread out radially to the different hairs of the circle which they supply (Plate 3, Fig. 12, for. pi’ph.). There are three questionable points to be settled in regard to the innervation of the otocyst, and the same is true for the sensory bristles of decapod Crustacea in general. a. Is each hair supplied by one nerve fibre and sensory cell, or by many ? b. How do the peripheral fibres terminate? Do they attach them- selves to a sense cell, or to some part of the hair, or do they end free? If this latter be the condition, does the fibre terminate at the base of the hair, or at its very tip ? c. Where do the fibres end in the central nerve organ, and how? For the determination of these questions, it is important to compare the conditions found in all kinds of sensory bristles. Because different types of hairs have been used in various Crustacea for the study of the nerve terminations, and this difference in kind of material employed by various investigators may account for the very diverse conclusions they have drawn. All sensory bristles of decapod Crustacea can be divided into two general types : (1) Tactile bristles (Plate 2, Fig. 8) have typically a long, straight, plumed, attenuate shaft, attached at the base by a thin spherical en- largement, which allows great freedom of movement. Auditory hairs, so called, are merely modifications of these, for all gradations between the two exist. Tactile hairs are found on nearly all the appendages, and on some parts of the body. (2) Olfactory bristles (Plate 4, Fig. 13, set. olf., and Fig. 14) are short, cylindrical, or slightly tapering, and firmly attached as compared with tactile hairs, there being no marked basal enlargement. At the tip, the 186 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. } chitin is either pierced by a pore, or ends in a thin permeable membrane, which allows substances in solution to enter the cavity of the hair. It found on the first or second antennex, they are termed olfactory hairs; _ when on the oral appendages, ¢aste or gustatory bristles, though thei is functions are probably the same. . a. Number of Nerve Elements toa Single Bristle. Until 1891 it was supposed that only a single ganglion cell and fibre-process supplied each hair, Then Vom Rath (91, p. 207) asserted, that beneath every sensory — hair of crustaceans there is a large group of ganglion cells, each sending out a peripheral process, these converging and entering the base of the hair as a single large strand. This opinion he again expressed in 1894 foal all arthropods. He did not study the innervation of the otocyst, but — apparently confined his attention to the olfactory type of hair, as his — figures are all of unfringed bristles. The number of elements supplying each hair of the otocyst can be determined by, first, counting the number of fibres in the auditory nerve, | and the number of nerve cells connected with these fibres, and then, secondly, comparing the statistics thus obtained with the number of hairs in the otocyst. If there is but a single cell and fibre to a hair, these numbers should coincide, at least approximately. But if there are always numerous elements, as Vom Rath maintains, then the number of fibres and nerve cells should be many times that of the hairs. The number of fibres can be readily counted in a transverse section of the — otocyst nerve stained mtensely with iron hematoxylin and only slightly decolorized. The ganglion cells can be enumerated in serial sections cut in the plane of the long axes of the cells, so that their characteristic size and bipolar condition (seen in Plate 2, Fig. 6) will readily distinguish them from the hypodermal or matrix cells. From many such counts, the number of nerve elements was found to be approximately equal to that of the hairs. For example, in one otocyst there were 55 hairs, 53 fibres in the nerve supplying them, and 58 cells connected with these. The number of cells could not be determined with perfect accuracy, as some cells may have been halved in sectioning. Slight variations in the numbers, however, are not of great significance, as, in order to have even two nerve elements to a hair, the number of fibres or cells must be at least twice as large as that of the hairs. Moreover, the ganglion cells are always isolated, and each is surrounded by a separate sheath ; their fibres are also separated from each other. Neither cells nor fibres occur in groups surrounded by a common sheath as Vom Rath (’92) deseribes them. Jn the otocyst, then, there is but one nerve element to each hair. — PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 187 In the tactile hairs the same methods of procedure were followed ; and further evidence was obtained from methylen-blue preparations. One of these is shown in Figure 11 (Plate 3). It will be observed at once from this figure that there is only one cell and one fibre to each hair. But in other preparations of the same appendage (Plate 4, Fig. 14) from two to ten cells are found grouped together irregularly, and sending all their processes to the same bristle. When this was the case, 7¢ was always observed, that the hair so supplied was of the short, blunt, fringeless type, and so possibly not a tactile but an olfactory hair. So far, the evidence has been entirely against Vom Rath’s statement ; but if we examine the innervation of the olfactory bristles, entirely differ- ent conditions will be found to exist, and in complete accord with his conclusions. On the inner flagellum of the first antenna of Palemonetes numerous olfactory bristles are found, arranged in rows of four or five hairs each (Plate 4, Fig. 13). The nerve cells and fibres supplying these hairs stain beautifully with methylen blue. Only single elements at first appear, but if the stain is allowed to act for a longer period, nearly every cell and fibre will become impregnated. It can then be seen that a large number of elements supply each hair. The cells are packed so closely together as to make the counting of a group difficult, but many counts upon sections stained in hematoxylin make it certain that more than a hundred cells may compose a single group, and supply a single olfactory hair. The cells send off each a peripheral fibre. These fibres enter the base of an olfactory hair as a single large strand, 12 to 15 w in diameter. In Figure 13 only a few of the elements are shown; the sheath, which surrounds both cells and fibres, marks the outline of the spindle-shaped group of cells, and shows the size of the fibre strand. The gustatory hairs on the oral appendages are also each supplied with numerous nerve elements (Plate 4, Fig. 14). The number is not nearly so great as in the olfactory hairs, —averaging about 10 to a hair, —nor are they so regularly and compactly grouped. They differ markedly, however, from the conditions found in tactile and otocyst hairs. The distinctly different conditions —as regards the number of nerve elements of the hairs — found in the olfactory and otocyst bristles, seem to explain the diverse conclusions of Bethe and Retzins on the one hand, and Vom Rath on the other. The two former observers worked on the tactile type of sensory bristles, while Vom Rath, as his figures show, evidently confined his attention to the other type. The conditions which 188 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Vom Rath found in the olfactory type he too hastily attributed to all the sensory hairs of Crustacea. . b. Peripheral Terminations. Here again we find a difference of opin-— ion. Hensen (63, p. 368) asserted that the peripheral fibre was attached to a process (lingula) from the base of the hair shaft. Claus (91), Vom Rath (92, ’94), and Bethe (’95) found fibres reaching to the | very tip of the sensory bristles; while Retzius (’95, p. 17) found no~ evidence of nerve terminations beyond the enlargement at the base of © the hair in decapods, though he observed in Entomostraca the same con- ditions as did the other three investigators. I have obtained hundreds of preparations of nerve endings in the various sensory hairs of Palemonetes with several of the best modern nerve methods, and all furnished the same evidence. The conditions found for otocyst hairs were in every case as illustrated in Figures 4_ and 8 (Plates 1,2). The ganglion cells, as already noted, lie at some distance (0.25 to 0.40 mm.) from the bases of the hairs which they supply. The reason for this becomes obvious, when the formation of the new hairs is considered. The developing hair tube extends below the base of — the old hair a distance equal to at least one-third the length of the hair, and the ganglion cells necessarily lie below the lower or proximal end of the hair tube (Plate 3, Fig. 10, tb. set.). Hence they must be at least a third the length of the hair distant from its base, though they occupy a closer position directly after ecdysis than for some time before. The terminal fibres (Plate 2, Fig. 8, fbr. n.), which are as long as the dis- — tance of their cells from the hairs, enlarge slightly as they near their termination, and always end in the expanded base of the hair directly — below the shaft proper. There are no signs of attachment to any part of the wall of the hair, nor of fine branching of the distal end of the fibre, such as Retzius (’90) describes. Figure 4 (Plate 1) shows diagram- — matically one nerve element of the otocyst, the position of the ganglion cell, and the ending of its peripheral fibre in the base of the hair. In Figure 8 (Plate 2) only the termination of the fibre, highly magnified, is given. The elements of the tactile hairs end in precisely the same manner as those of the otocyst. A number of these endings are shown in Figure 11 (Plate 3). In no case was a nerve ending demonstrated in the shaft of the hair. Thus, all the evidence of preparations goes to prove that in both otocyst hairs and tactile hairs the nerve fibre, without branching, ends in the enlargement at the base of the hair, and never enters the shaft itself. Se = See mi a Sa let a" PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 189 In the olfactory bristles the cells are situated about 0.40 mm. posterior to the bases of the hairs, and their peripheral nerve fibres, stained by methylen blue, were traced tn almost every preparation, some distance into the shafts, though in the tactile hairs of the same appendage no fibres could be followed further than the base. Figure 13 (Plate 4) shows the olfactory endings, some of them extending half the length of the hair shaft, but none as far as the tip; nor was such a condition ever found, although a great number of preparations were examined. The direct evidence of preparations shows, then, that the peripheral nerve endings are different for the different types of hairs. The fibres terminate in the enlarged base of tactile bristles, while in olfactory hairs they end Sree in the shaft itself. This direct evidence is strengthened by other structural conditions. (1) Owing to the rigidity of the hair shaft and its delicate basal attachment, a mechanical stimulus applied to a tactile hair would be apt to produce its strongest effect at the base. Therefore we should expect to find the nerve termination at this, the point of greatest stimu- lation. The innervation of the tactile hairs of vertebrates extends only to the base, yet the slightest touch of the hair tip stimulates the nerve ending. Similarly, in the otocyst hairs the point of greatest stimulation must be at the base. The hair tips are so entangled with each other, and with the otoliths resting upon them, that a stimulus applied to one must affect them all. If this stimulus is caused by the shifting of the weight of the otoliths resulting from a change in the direction of the pull of gravity, it will affect the delicate, labile articular membrane at the base of the hairs far more vigorously than the part of the shaft attached to an otolith, or entangled with the tip of another hair which is so attached. In the olfactory hair, on the other hand, the chemical stimulus finds access through the permeable tip, and, traversing the cavity of the shaft, comes at once into contact with the terminations of the nerve, which here, as we have seen, runs some distance toward the tip of the hair. This, then, isa condition of affairs which, in view of the function of the olfactory hairs, we should reasonably expect. (2) The conditions during hair formation are very unfavorable to the assumption that the nerve fibres extend to the tips of the tactile and auditory hairs. In adult Palamonetes, a month at least before ecdysis takes place, the matrix cells withdraw their processes to the basal por- tion of the hair, leaving the upper part of the shaft empty. As the j ™ 7 190 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. shrimp moults once in two or three months, this means that for nearly half the time the nerve fibre cannot extend further than the base of the hair. Yet the animals are apparently as sensitive to stimuli during this period as at any other. After the new hair is fully formed, and its tip projects into the base of the old hair, which has now lost all direet nerve connection, the animals still respond quickly to tactile stimulus ; the impulse resulting from the stimulus is transmitted from the tip of the old hair to its base, thence to the shaft of the new hair, by which in turn it is transferred to the nerve fibre. (5) If certain of the nerve fibres supplying the tactile hairs are stained with methylen blue just before ecdysis when the new hairs are fully formed but still deeply invaginated (Plate 3, Fig. 10, ¢b. set.), they may be traced some distance into the shaft of the new hair. Now, by removing with a fine needle the old test, cta., the new hairs can be pulled out into their functional position. The nerve fibres, however, are not pulled out with the hair the whole distance, but remain nearly in their original relative positions, barely projecting into the bases of the hairs, a condition already pointed out in Figure 11 (Plate 3). It is unfortunate that the investigators of these nerve endings have never taken into account the tissue changes — certainly of great impor- tance — which occur in all Crustacea between moults. At certain stages in their formation the delicate protoplasmic pro- cesses in the tips of the new hairs stain very sharply, and have a varicose appearance, similar to that of nerve fibres; as these project some distance into the old hairs, they might easily be mistaken for terminal nerve endings. c. Central Terminations. By means of methylen-blue preparations the nerve fibres supplying the otocyst were traced continuously in their course from the sac to their central endings. Whole preparations of the anten- nules and brain could be used for this purpose, as the tissues were ex- tremely transparent. On account of the proximity of brain and otocyst, the nerve supplying the latter is very short. It enters the anterior end of the brain lateral to the antennular nerve, the two joining as they pass within (Plate 3, Fig. 12). While the antennular nerve pursues a straight course, the other (Figs. 2, 4) descends from the sensory hairs in the floor of the otocyst, forms the sensory ganglion, and in continuing its course approaches somewhat the median plane and describes the form of an elongated letter S, the plane of which is dorso-ventral. Just before the two nerves unite to enter the brain, a third smaller 7 PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 191 nerve is received by the otocyst nerve on its dorsal] side (Plate 1, Fig. 2, rm. /.). This nerve is formed by an aggregation of fibres from the tactile bristles of this segment of the antennule, and runs almost straight toward the median plane till it joins the nerve of the otocyst. The fibres of the latter enter the anterior end of the brain ventral to the optic neuropil, and median to the globulus (Plates 1, 3, Figs. 4, 12); they extend backward to near the posterior end of the central organ in an almost horizontal plane, lateral to the fibres of the antennular nerve. They end in a region just anterior and median to the neuropils of the second antenne, branching into delicate dendritic fibrille, which form a well-marked neuropilar mass (Fig. 12, for’.). Fibres supplying the tactile hairs of the basal segment of the anten- nule end in the same neuropil, while the main nerve to the antennule ends in a closely connected fibrillar mass just median to it. No nerve cells were found in the brain connected with the sensory fibres from the otocyst. Association elements, with large dendritic branches, put these neuropils into communication with the optic centres. One of these con- necting fibres is shown in Figure 12 (for. ass.). Its cell, which sup- posably exists, was not stained. According to Bethe’s (’97, Taf. xxviii. an.1 ) experimental work on the brain of Carcinus manas some of the otocyst fibres should end in the globuli. He could not demonstrate such fibres, however, in his preparations of the crab’s brain, nor was I able to obtain conclusive evidence of such endings in the globuli of Palzemonetes. d. Histology of the Nerve Elements. The nerve fibres of Pale- monetes are relatively large ; those of the otocyst reach their greatest size immediately before they enter the neuropil substance of the brain. At that point in their course they are from 3 to 5 w in diameter, not including the nerve sheath. In a transverse section of the nerve the sep- arate fibres show distinctly, as they are held apart by connective tissue. The fibrillar structure was made out definitely only in methylen-blue preparations which had been well differentiated in process of fixation. The gold-chloride method of Apathy, though tried several times, did not give a successful reaction. Fibrille were made out distinctly in only one preparation, though some evidences of such structure appeared in many. Figure 15 (Plate 4) shows a portion of a peripheral fibre in which many fibrils are seen running longitudinally. No single fibril was traced any considerable distance, nor could any evidence of the fibrils be found in the ganglion cells. The fibrillea are embedded in a 192 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. semi-fluid, homogeneous substance, which is the first to take up the methylen-blue stain. It has been called by Bethe (98) the “ peri- fibrillar substance.” The accumulation of this fluid into drops gives the characteristic beaded appearance of methylen-blue preparations. A distinct nucleated myelin sheath surrounds both the fibre and the peripheral ganglion cells of Palazmonetes. This sheath, which stains intensely black in Vom Rath’s platino-osmic fixative, can be traced some distance beyond the peripheral ganglion cells toward the sensory hairs, and also centrally into the brain, where it ceases only when the fibres enter the neuropil substance. Figure 16 (Plate 4) shows a ganglion cell and its peripheral process surrounded by the sheath. Elongated, flattened nuclei occur at intervals along the walls of the sheath, curved around it and the enclosed fibre; certain of these sheath nuclei can be seen in Figure 4 (nl. tu.) between the ganglionic cells and the brain, though the myelin sheaths are not stained in this hematoxylin preparation. Quite frequently one of them may occur in close proximity to a ganglion cell. Thus are produced (Plate 4, Fig. 17) appearances which might be mistaken for a ganglion cell with two nuclei. Careful study, however, shows that one nucleus (z/.) lies within the cell, the other (n/. fu.) without, but abutting on the ganglion cell so closely as to sometimes change its form. In every instance of this kind one of the nuclei, owing to its irregular outline, its smaller size, and the curved form which it takes in adaptation to the surface of the cell, could be identified as belonging to the sheath rather than to the nerve cell. The peripheral ganglion cells are much elongated and are of the typical bipolar form (Plate 4, Fig. 18). They measnre from 10 to 14, in diameter; their nuclei are relatively large, measuring from 7 to 9 in diameter, and are usually ovate in outline, their length in some cases being twice as great as their diameter. One large spherical nucleolus is usually present in the chromatic network, though some- times two or more are found. No definite structure can be recognized in the cytoplasm of the cell, nor any traces of fibrille ; this, however, is not strange, as the cell usually stains so intensely that it would not be reasonable to expect to make out its finer structure. In methylen- blue preparations a narrow zone about the nucleus stains only faintly, the coloration becoming more intense as the periphery of the cell is approached ; so here, as Bethe also found in the nerve cells of Carcinas, the chromatin granules are more numerous at the periphery of the cell cytoplasm, and nearly wanting around the nucleus. — PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 193 3. Development of the Otocyst (in Homarus americanus Milne-Edwards). In order that the development of the otocyst in the lobster may be more readily understood, it may be best to compare briefly its adult condition with that of Palzmonetes. : It was dissected and described by Farre (’43), and again by Hensen (63). The sac is drawn out posteriorly into a dorso-ventrally flattened projection, the “cochlea” of Hensen. The external aperture is extremely small, guarded by bristles, and located at the median, dorsal, and ante- rior end of the sac, the dorsal wall of which, like the dorsal wall of the antennule, is very thin, forming the so-called tympanic membrane. On the floor, which is nearly horizontal, there is a semi-circular ridge - (Plate 5, Figs. 24, 26), which forms the sensory cushion. From this arise the otolith hairs, which have straight shafts, and number from 500 to 600. The four rows of these are so arranged as to form a semi- circle, the open side of which (at the right in Plate 5, Fig. 26), is ante- rior instead of posterior as in Palemonetes. At the anterior end of the curve there is an irregular group of smaller hairs, with bent shafts. On the median wall of the sac, near its posterior end, there is an irregular double row of long thread-like hairs, with shafts heavily fringed (Fig. 26, set. m.). The otoliths are numerous, and rest on the area surrounded by the rows of sensory hairs, and also on the hairs themselves; the thread-like hairs are free, and float out into the lumen of the sac. Not much has been written on the development of the otocyst in decapods. Reichenbach (86), in his work on the embryology of the crayfish, figures the invagination of the “auditory sac” at an early stage in the egg. The crayfish, however, as it develops into the adult form without passing through the larval stages characteristic of most other decapods, is not a typical example. Herrick (95, p. 194) alludes to the appearance of the otocyst cavity in the third larval stage of Homarus, and he shows its position at this stage in connection with the development of the first antenna. In the fourth stage it is a shallow depression containing a few otoliths and in the fifth larva its aperture begins to close. I shall describe its condition in the first four larval stages. a. First Larval Stage. (Schizopod larva, without abdominal appendages.) Sections of lobster eggs in different stages up to time of hatching showed no evidence of the otocyst in the antennule, and it became 194 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. apparent that its development took place wholly in the free-swimming stages. A transverse section through the antennule of a newly hatched larva (Plate 4, Fig. 19) shows no sign of invagination in the region where the sac is to appear. But certain elongated nuclei, evidently those of modified hypodermal cells, are found grouped, two or three layers deep, beneath the dorso-lateral wall of the appendage (Fig. 19, cl. ma.). These elongated nuclei, viewed from the dorsal surface of the appendage, are seen to be roughly arranged in a semi-circle, like the rows of otocyst hairs in Figure 26 (Plate 5), and when traced through later stages, the position they occupy is found to be directly beneath the ridge where the sensory hairs later appear (Plate 5, Fig. 24, set. ot.). They are evidently, therefore, the nuclei of the matrix cells which build up~ by secretion the chitinous walls of the sensory hairs. These cells, like those which take part in hair formation after ecdysis, originate from the chitinogenous hypodermal cells by simply becoming elongated and sink- ing beneath them. A similar arrangement of matrix cells was found in the developing otocyst of Mysis by Bethe ('95*). Numerous spherical nuclei, which stain in a manner characteristic of nerve cells, are present just below the matrix cells (Fig. 19, ’6l.). If traced back to the gan- glionic masses of the brain, they are found to be continuous with the nerve cells of the latter, and probably originate from them. b. Second Larval Stage. (Second to fifth pair of abdominal appendages present.) In this larva the first evidence of invagination is seen on the dorsal side at the base of the antennule (Plate 4, Fig. 20). The nuclei of the matrix cells are now larger, and very conspicuous at the lateral side of the transverse section, the region where the rows of hairs will later appear. Figure 22 (Plate 4) shows the anterior and posterior limits of the invagination and the fundament of the sensory ridge, marked by a fold in the hypodermis and chitin at cl. ma. The matrix cells just posterior to this fold, whose processes are directed toward it, are those which are to form the transverse portion of the hair rows. As in the first stage, nuclei of nerve cells lie immediately beneath the matrix cells, but the cytoplasm about them shows as yet no definite boundaries or outlines, nor are there any signs of nerve fibres connected with them. ec. Third Larval Stage. (Chele relatively larger, uropods present.) In this stage (Plates 4, 5, Figs. 21, 23) invagination has proceeded PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 195 still further. There is a deep lateral, as well as a posterior, fold in the chitin ; but the sac, if it can now be called such, is very shallow, wide- mouthed, and without sensory hairs or otoliths. From the group of matrix cells, however, the tips of embryonic sensory hairs‘may be made out, projecting dorsally, but covered by the chitinous floor of the sac (Plate 5, Fig. 27). Only after the wall of the sac has been shed at the next moult will they become functional organs. d. Fourth Larval Stage. (Form like that of adult ; thoracic exopods rudimentary.) The sac has now greatly increased in size, and nearly tills the cavity of the appendage (Figs. 24, 25). Its opening has become smaller, and is protected by numerous fringed bristles, which project from its sides (Fig. 25, tct.). About 200 sensory hairs are present borne on a prominent sensory ridge (Fig. 24, set. of.) and arranged in three regular rows, one row less than in the adult stage (Fig. 26). The whole band bears some resemblance to a sickle. Beginning at the median side of the sac floor, the rows curving only slightly run laterally, then with a Stronger bend turn forward. At the anterior end of the sac regular arrangement ceases, the hairs being grouped promiscunously. Besides these large hairs on the sensory ridge, which measure 120 uw to 150 u in length and from 4 u to 6 u in diameter, there is, as in adults, an irregular row of more attenuate hairs arranged longitudinally along the posterior part of the median wall (set. m., Fig. 26). They number about thirty, are ou the average 140 w in length, and have a diameter of only 2uto3u at the base of the shaft. Many otoliths, consisting of fine particles of sand, rest on the hairs of the sensory ridge, as in the adult condition, but do not come into con- tact with the attenuate bristles of the median side-wall, which project free into the liquid contents of the otocyst. The sensory ridge is much more prominent at this stage than in the adult. This, and the size of the aperture, are the chief differences between the two, and are well shown in Figure 25. The opening gradually becomes smaller in the fifth, sixth, and seventh stages, until in the full-grown animal it is almost obliterated. A fourth row of hairs, not yet developed, is formed posterior to the others at some stage later than the seventh moult, this being the oldest Stage that I have studied. Except for the gradual closure of the aper- ture, the larve of the fifth, sixth, and seventh stages show the same conditions in the otocyst as the stage under consideration. In Figure 24 (Plate 5) ganglion cells (cl. gn.) are seen beneath the 196 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. sensory ridge. The origin of these could not with certainty be traced out in the material at command, though from the conditions found jn the first stage, it is probable that they are derived from the neuroblast cells of the brain. ‘The only evidence in favor of this view is the prox- imity of the brain, and the fact that at an early stage nerve cells which were continuous with the ganglionic masses of the brain were present beneath the matrix cells of the otocyst. F igure 26 shows, somewhat diagrammatically, the general innervation of the otocyst hairs of the | fourth larval stage, as brought out by methylen blue. The condition is essentially that of the adult. There is but one nerve element to each hair, and the endings are-in the enlarged bases. No myelin sheath is developed in either the larva or adult lobster. Central terminations of the otocyst fibres were not traced out, nor was their finer histology — investigated. The most striking point to be noted in the development of the otocyst of the lobster is the abrupt change which takes place after the third meult. The shallow, functionless depression of the third stage is con= verted at once into the active, well-differentiated organ of the fourth larva. This sudden leap in the development of the otocyst is correlated with an abrupt metamorphosis of the larva’s general form and method of locomotion. As this correlation may have an important physiological significance, it will be discussed in detail in the theoretical portion-of this paper. II. CranGon VULGARIS Say. 1. Structure of the Otocyst. a. Sac. The otocyst has been described only briefly by Hensen (’63), He figures the sac dissected out, and gives two sketches of the sensory, hairs, and the prominence upon which they are borne. The sac, as seen in a section passing through its middle and trans- verse to the long axis of the antennule, has the form of a half-circle. In a cross-section more posterior its outline is made irregular by the pro- jection of the sensory ridge or cushion from its lateral wall (Plate 6, Fig, 28). This is an entirely different condition from that found in Pali monetes, where the sensory cushion is basal. More irregular still is its form in frontal section, as shown at ers. sns. in Figure 29 (Plate 6). The dimensions of the sac in individuals of medium size (25 mm. long) are: length 0.44 to 0.55 mm. width 0.28 “0.38 ‘ (anterior to sensory ridge) depth 0.20 “0.22 « ; lla ep eT te ee wg ation ae ets an 2 ee. gta PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 197 It is thus relatively wider, and more shallow than that of Palemonetes. The wall is of thin chitin continuous at the large oval aperture (Plate 7, Fig. 30) with that of the dorsal side of the antennule. The aperture is as wide and nearly as long as the sac itself; instead of a fold of chitin it has for protection a row of large fringed bristles. These are ranged close together along the posterior edge of the opening and extend their long parallel shafts beyond its anterior margin. A fine-meshed grating is thus formed, through which even microscopic organisms could not pass without displacement of the bristles. b. The sensory cushion (Plate 6, Fig. 29, ers. sns.), as already noted, projects from the posterior portion of the lateral wall of the sac. Its direction is not transverse to the long axis of the sac, but it points obliquely forward and mediad. It is a ridge rather than a cushion, for the hairs are arranged in a short, nearly straight single row, instead of in several rows having the form of a sickle. This row of hairs, which defines the limits of the sensory region, starting at the dorsal end of the ridge, takes a course along its convex surface downward and backward, and ends where the ridge disappears, just before the floor of the sac is reached. A portion of a row of hairs is shown in the right otocyst, Figure 29, set. ot. (Plate 6), where the hairs anterior in position are really above or dorsal to those posterior to them. The ridge-like projection of the sensory prominence is best seen in a parasagittal section (Plate 7, Fig. 30, set. of.), a hair being there shown at the apex of the ridge. The matrix cells are essentially the same as in the hairs of Palemonetes. They occupy the region just beneath the bristles, into which their processes extend. The space in the sensory prominence below and lateral to the matrix cells is occupied by the sensory ganglion cells, the fibres from which penetrate between the formative cells and reach the bases of the hairs (Fig. 29, el. gn). c. Structure of hairs. Arranged on the sensory ridge in the manner above described, the hairs of the otocyst are 26 in number, as shown by the average of a large number of individuals. They are largest at the upper anterior end of the row, where they measure 180 uw in length and about 94 in diameter at the base of the shaft. Proceeding down the line they are successively smaller, the last of the series being only 100 u in length and 6 in diameter. There is a conspicuous spherical enlarge- ment at the base of the hair shaft (Plate 7, Fig. 31, md. sph.), as in the otocyst hairs of Palemonetes. The shaft itself for about a third of its length projects straight out horizontally into the lumen of the sac. Then it bends down ventrally nearly at right angles, though the amount VOL. XXXVI.— No. 7 8 198 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. of curvature is different for different hairs. The larger, being in a more elevated position, usually bend at a sharper angle than those near the floor of the sac. All are heavily plumed; the pinnules are long and_ | coarse (Plate 7, Fig. 31, penn.) and often have otoliths firmly attached — to them by a substance probably of glandular origin. Hensen (’63) describes the ctolith hairs of Crangon, as follows: “ Es steht namlich auf die schon erwahnten Vorbuchtung eine einzige Reihe von 7 oder 8 — Haaren; diese Haare reichen bis zur Kugel in die Steine hinein, chre Zahl erscheint viel zu gering fiir deren Masse. . . . Sie sind 0.075 mm. lang, 0.0075 mm. breit und gerade anfgerichtet.” This description of these hairs is completely at variance with the conditions I have found in the American Crangon. In order to deter- mine, therefore, whether this was a true specific difference, or due to an error on Hensen’s part, a number of the European specimens, procured by Dr. Mark from Professor Herdman in Liverpool, were examined. After dissecting ont the otocysts of 12 specimens, I was entirely satisfied that Hensen’s description was incorrect. The hairs are precisely the same in size, form, and number as in the American variety. They have their shafts distinctly bent near the tip at angles varying from 25° to 90°; of the individuals examined none possessed less than twenty-four hairs in the sac, the average being twenty-six. That Hensen should have made such a mistake is not strange. He | himself says: “ their number appears much too small for the mass [of the otoliths].” The tips of the hairs are concealed by the otoliths, and only the first third of the row would be visible from above. d. The formation of hairs after ecdysis is identical with that of Palzemonetes. e. The otoliths are numerous, larger than in Paleemonetes, and found mostly in the posterior part of the sac, in contact with, or even attached to, the fringed tips of the hairs. Mainly siliceous, they are taken in after each moult, being readily pushed into the large opening of the otocyst. They can be almost completely washed out by a fine jet of water introduced artificially, and if the animal so treated is then placed in an aquarium containing iron filings, or other substitute, this material will soon be used to replace the otoliths of sand. 2. Innervation of the Otocyst. As in Paleemonetes, the brain is very close to the otocyst, and the nerve supplying the sac is therefore short. Its general course is shown at n. of. in Figure 29 (Plate 6). PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 199 Leaving the anterior end of the brain with a bend away from the median plane, it gives off in front of the globulus a small lateral branch (rm. 1.), which supplies the tactile bristles of the antennule. The main nerve, after passing between the globulus and the posterior end of the sac, runs forward only a short distance to the sensory prominence on the lateral side of which its ganglion lies. ‘The peripheral fibres can be traced forward and slightly mediad from the ganglion to the bases of the otocyst hairs. The whole course of the nerve is approximately in a frontal plane, though its peripheral ending is slightly more ventral than its point of departure from the central organ. In Figure 28 (Plate 6) the transverse section of the antennular nerve (n. at.1) is seen to be median to the sac, while the ganglion cells of the otocyst nerve (cl. gn.) are lateral to it. a. Number of Nerve Elements to a Single Bristle. There is in Crangon bnt one ganglion cell and fibre to each otocyst hair. The cells and fibres were counted as in Palemonetes, and the numbers thus obtained were found to agree approximately with the number of the hairs. Methylen-blue preparations of the olfactory nerve elements were obtained, and the conditions there brought out agreed essentially with those found in the same type of hair in Palemonetes, large groups of nerve cells being present beneath each olfactory bristle. b. Peripheral Terminations. Nerve fibres to otocyst hairs were never traced beyond the enlarged base of the bristle, where they end free without branching. A typical nerve element of the otocyst is given diagrammatically in Figure 29; it shows the peripheral ending of the fibre at the base of set. ot. In the olfactory hairs, on the other hand, the nerve fibres in most cases could be traced up into the shaft of the hair, though never through its whole length. Thus in Crangon, as in Palemonetes, there is a distinct difference in the innervation of the two types of bristles, both as to the number of elements, and in the manner in which the fibres end. ce. Central Terminations. Centrally the otocyst nerve ends in a posi- tion (Fig. 29) corresponding to that of the central terminations in Pal- emonetes, but the fine fibrillar branching, which was brought out distinctly by methylen blue in that form, could not be impregnated in Crangon. d. Histology of the Nerve Elements. So far as worked out, this was similar to that already described in Palemonetes. A myelin sheath is present in Crangon as well as Palemonetes, though it was not observed in any other decapods. 200 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 3. Development of the Otocyst. This was not studied in Crangon. III. Camparus aFFInis (Say) GrRarp. The otocyst of the crayfish has been figured by only Farre (’43) and Huxley (80). The description of the former investigator was excellent for the time at which it was made. Huxley alludes to the otocyst in his work on the crayfish, and gives one figure showing the sensory region dissected out. Hensen (’63) describes the hairs of the otocyst in Astacus fluviatalis, but does not touch upon its other structures. 1. Structure of the Otocyst. a, Sac. The otocyst of Cambarus (Plate 8, Figs. 37, 38), except for its smaller size, resembles that of the lobster very closely. The aper- ture, exceedingly small in the lobster, is here quite large, though, on account of the dense chevaux de frise of fringed bristles, it seems smaller than it really is. These bristles, projecting from around its margin, effectually cover and conceal the opening. It occupies the middle of the dorsal side of the antennule; its anterior margin corresponds to the anterior wall of the otocyst, and it extends back from this point nearly one-half the length of the sac. Its width is about one-third that of the otocyst (Fig. 37). The cyst does not by any means fill the cavity of the antennule. It is rounded off in front, but sharply pointed at its posterior end, where it is very shallow (Fig. 38). Its walls are of uncalcified chitin and continuous with the very thick calcified shell of the antennule (Figs. 37, 38). Its dimensions in average-sized animals are : Length from 1.75 mm. to 2.25 mm. Width 4° (1.208 ORO Depth ot) 0:85) ir eel Aes b. Sensory Cushion. The sensory ridge, or cushion, in the base of the otocyst is not prominent, as that part of the sac floor upon which the sensory hairs are borne is but slightly elevated above the rest (Fig. 38, set. ot.), and, contrary to the conditions found in the two forms already described, the sensory surface is nearly horizontal, instead of being vertical or oblique. The arrangement of the hairs is shown in Figure 40 (Plate 8). Three sets can be distinguished, corresponding to the divisions of the otic nerve, —a median, a lateral, and a transverse or posterior. The first and third are nearly straight, the second sickle- PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 201 shaped. The “median” set consists of a single nearly straight row, running from the posterior angle of the sac obliquely forward and mediad, back of which there are two or three shorter, irregular rows of scattered hairs. The lateral set consists of two concentric rows, which have the form of a crescent or the blade of a sickle, the handle of which is represented roughly by the nerve trunk connecting ‘the bristles with the brain. The hairs of the outer row are much larger than those of the inner series. At the tip of the sickle blade the area covered by the bristles expands, and the hairs are arranged in 4 or 5 irregular rows. Behind the proximal end of this sickle-shaped double row of bristles is a short row of very large hairs, the posterior set (Fig. 40, set. p.), usually nine in number, which extends transversely across the posterior portion of the sac immediately in front of its pointed base. Matrix cells are found in the region directly beneath the hairs, as in the other forms described (Plate 8, Fig. 37), and the nerve cells with their peripheral fibres lie below the chitin, either just within (lateral set), or slightly posterior to (median and transverse sets) the rows of hairs (Plate 8, Fig. 40). By looking down upon the floor of the sac one can make out numerous small pores (represented in Figure 40 by minute circles), which penetrate the chitinous wall in that portion of the floor which is inclosed by the sensory bristles, especially in its lateral part. In transverse sections some of these pores are cut through, and it then appears that they connect with the ducts of multicellular glands which are located in the tissues beneath. One of these glands with its duct and pore is shown in Figure 39. It is apparently similar to the tegu- mental glands found in different parts of the lobster and figured by Her- rick (95, Cut 5, p. 77). In Cambarus these glands evidently supply the secretion which attaches the otoliths to the pinnules of the otocyst hairs. e. Structure of Hairs. This has been described in some detail by Hensen (63), to whose descriptions I have not much to add. The hairs are very similar in structure to those of the lobster. Their number varies greatly in different individuals, but is usually over 200. The straight, or only slightly curved, shaft is heavily fringed, and borne on the customary spherical base. Their dimensions are : Length, from 65 mw to 175 p. Diameter, “ 154“ 18 yp. A transverse section of the shaft near its base has the peculiar shape Shown in Figure 35 (Plate 7). This modification of the form of its wall, found also in the otocyst hairs of the lobster, doubtless renders the shaft more rigid than if it were a simple hollow cylinder. bo 02 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. The shaft, as already noted, is nearly straight, but it is attached to the floor of the sac in such a way as to make a very small angle with its surface, being, in fact, nearly parallel to it. Thus in Cambarus the bending has taken place at the base, not, as in Palemonetes and Crangon, in the shaft itself. In these two forms the tendency of the shaft to bend must be aided, if not caused, by the weight of the otoliths attached to the slender tips of the hairs. In the lobster and crayfish the modified form of the shaft makes it too rigid to thus give way, and the bending, if any, must take place at the thin, membranous basal sphere. d. Formation of Hairs. (Not studied in Cambarus.) e. Otoliths. These are composed of large grains of sand distributed mostly within the circle of hairs, and supported in part by them. As the sac has a large opening, they are readily taken in through it after each ecdysis. 2. Innervation of the Otocyst. As the crayfish was well adapted for work with methylen blue, a large number of preparations of the sensory nerve elements were made, not only of the hairs of the otocyst, but also of the other sensory bristles. The nerve supplying the otocyst issues from the ventral surface, instead of the anterior end, of the brain, and at once passes forward with a slight lateral curvature to the pointed posterior end of the sac, beneath which its fibres spread out to the different hairs. It divides roughly into two strands, one of which passes obliquely forward and mediad to supply the median set of bristles (Plate 8, Fig. 40), while the other follows the course of the lateral sickle-shaped set, lying on the concave side of the two rows, to which it gives off fibres along its whole course. Before this division of the nerve takes place, a few large fibres run out from it on the lateral side (Fig. 40) to supply the short transverse row of large bristles (Plate 7, Fig. 33). The sensory nerve cells lie immediately beneath the hypodermis, and their peripheral fibres run in a plane parallel with the floor of the sae. In the case of the transverse rowof large hairs, the nerve cells are situated about 450 w posterior to the bases of the shafts, their peripheral fibres being therefore nearly half a millimetre in length. This is accounted for by the position of the new hair tube during the period of its formation between moults, when it extends back from the base of the functional shaft 350 1; the distance from base of hair to ganglion cell must consequently be somewhat greater than this. PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 203 a. Number of Nerve Elements to a Single Bristle. The number of cells and fibres for the whole sac could not be determined with exact- ness, as other sensory elements, supplying tactile hairs, are mingled with those of the otocyst. But in the short transverse row of large hairs, the cells and fibres are sufficiently isolated to allow of their being counted in serial sections. There are but nine hairs in that row, and if the nerve elements supplying them were twice as numerous, it would be at once apparent. The cells always occur singly, and their fibres run separately and parallel with one another to the bases of their respective hairs (Plate 7, Fig. 33). The number of each was counted many times, and it is certain that the number of ganglion cells and peripheral fibres exactly equals the number of hairs. Whole preparations of these nerve elements stained with methylen blue gave regularly nine ganglion cells and fibres supplying the nine sensory hairs. In these few otocyst hairs, at least, there is, then, but a single nerve element supplying each. In the tactile hairs of the scaphognathite of the second maxilla, many methylen-blue impregnations gave conditions like that shown in Figure 34 (Plate 7), only one sensory nerve element being stained. In the short spike-shaped bristles found on this same appendage, from three to five ganglion cells (Plate 7, Fig. 32, el. gn.) were usually found sup- plying each bristle. In the olfactory bristles of the antennule, the conditions were the same as those already described and figured for Palemonetes, though fewer elements compose each spindle-shaped group of cells. b. Peripheral Terminations. No branching of peripheral nerve fibres was observed in any sensory elements, though many were traced the whole length of an appendage. In Cambarus the fibres end always at the base in the otocyst hairs (Plate 7, Fig. 33). There is often a marked increase in the diameter of the fibre near its termina- tion, caused either by the staining of its sheath at this point, or by a partial separation of the component fibrille. Tactile hairs show similar conditions in their nerve endings (Plate 7, Fig. 34). The fibre strands of the olfactory bristles were, on the contrary, traced into the shaft some distance, where they apparently end free. Thus in the crayfish, we have a distinct difference in the innervation of the two types of sensory hairs, which serves to confirm the statements made ecncerning the conditions in Palemonetes and Crangon. e. Central Terminations. The otocyst nerve in Cambarus is large enough to be dissected out and traced to the ventral side of the brain, which it enters lateral to the larger antennular nerve. Its point of en- 204 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. trance isa little to one side of the median plane of the brain, opposite the posterior end of the globulus (Plate 9, Fig. 41). Its fibres run backward and dorsad, just lateral to those of the ante:nular nerve, and end in a neuropil directly anterior and median to that of the second antenna (Fig. 41, m. ot.). The individual fibres end by branching into fine fibrillations, which could be traced only a short distance through the diffusely stained mass of fibrillar tissue about them. d. flistology of the Nerve Elements. The sensory nerve fibres of Cambarus are relatively smaller than those of Palemonetes. Imme- diately after leaving the ganglion cell each measures about 3m in diameter, but becomes smaller as it runs distally, until near the point of ending, where it again enlarges to its original size. In well differen- tiated methylen-blue stains, fibrillar structure is clearly brought out. Longitudinal sections, and whole preparations of continuous fibres, show fibrillations similar to those figured for Paleemonetes. The sensory nerve cells are relatively large ; they measure from 15 to 18 in diameter, and being bipolar are spindle-like in form. Their nuclei are spherical and from 10u4 to 12 in diameter. The cytoplasm of the cell never shows any evidence of fibrillations, but in methylen-blue impregnations there is a faintly staining zone directly about the nucleus ; the remainder of the cytoplasm takes on a deep blue color. This dif- ference in staining qualities may be due to the unequal distribution of chromatic substance in the cytoplasm. The myelin sheath, so characteristic for the nerve fibres of Paleemo- netes and Crangon, is not found in the nerve elements of the crayfish. 3. The development of the otocyst was not studied in Cambarus. Ac- cording to Reichenbach (’86) it is completely formed before the young animal leaves the egg. IV. Carcinus mM=nas Leacu. (Green crab.) We now come to the second type of otocyst, which is found in all brachyuran Crustacea ; it is closed, and without otoliths. Mistaken by Bate (’58) for an olfactory organ, and figured by him in the larval stages of the crab, it has been described carefully in Carcinas meenas by Hensen (’63) alone. His account, although fairly accurate, is in- fluenced by his seeing a fancied resemblance between the otocyst and the vertebrate ear ; the figures he gives of different parts of the sac dissected out leave one somewhat in the dark as to the relative positions of the structures described. PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 205 1. Structure of the Otocyst. a. Sac. The basal segment of the antennule in Carcinus is relatively large, and elongated laterally to such an extent that its width is nearly twice its length (Plate 9, Fig. 46). Along its dorsal wall there ex- tends transversely a distinct line dividing the chitin of the anterior part of the segment (/ad. a.) from that of the posterior. This line of division, which reaches from the lateral margin of the segment three-fourths of the way across its dorsal wall, is rendered more prominent from the fact that the chitin posterior to it (Jab. p.) is much lighter in color than that in front. If the antennule of a crab is examined directly after ecdysis, when the chitin is still very thin, soft, and uncalcified, this lighter colored area (Fig. 46, lab. p.) is found to be a fold, projecting forward over the ante- rior part ; and if its edge is lifted with a needle or fine pair of forceps, a transverse aperture is disclosed leading down to the lumen of the sac. This aperture extends from line 45 (Fig. 46) laterally down through the side wall of the antennule. There is, then, in fact, a free passage into the otocyst directly after moulting, a condition necessitated by the casting off of the old sac. But almost immediately after ecdysis, the opening is closed and its edges fuse together, probably owing to the simultaneous secretion of chitin by the hypodermis of the two surfaces which bound the orifice and are in direct contact. Figure 44 (Plate 9) shows at dab. p. the two surfaces which fuse. The form of the sac is very irregular, so much so that Hensen de- spaired of describing it. Its walls, like those of the forms already studied, are continuous dorsally with the calcified chitin of the antennule (Figs. 42-48, Plate 9). The sac is thus suspended from the dorsal wall 6f the appendage. Although composed largely of thin chitin, one portion of its wall is much thickened and calcified (mal., Figs. 43-48, Plate9). On account of its irregular outline measurements can be of only small value. The average of a number of measurements taken of the otocyst in speci- Mens approximating 30 mm. in length, gave the following results :— Greatest length, 1.11 mm. + (width, 1.96 “ fm idepth,1.0a.°% The seemingly contorted shape of the sac is caused by three protuber- ances or invaginations of its walls, which project into the lumen (Fig. 4, and Plate 10, Fig. 55). Two only of these prominences are sensory and bear bristles (Fig. A, set. ta. and set. fil.). The third and largest of 206 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. the three (mal.), which projects from the lateral and posterior wall of the cyst, is without sensory organs of any kind. Its wall is irregularly curved and pitted (Plate 9, Fig. 47 mal.,), while portions of it are even” calcified. At one point its walls are constricted to form a neck, which cars a large hammer-like head (Fig. 47). This is the “ Hammer” of Hensen, compared by him to the malleus of the vertebrate middle ear,” Figures 43, 48, and A show the relative position of this hammer to the _ a. Figure A. Model of the lumen of the left otocyst of Carcinus, dorsal view, tne upper wall of the sac removed. The cavity of the sac was modelled in wax from serial sections under a maguification of 50 diameters, and a plaster cast of the model photographed natural size In making the cut this was reduced to a magnification of 33 diameters. a., anterior; — m., median; set.’, group hairs; set. fil., thread hairs; set. ta., hook hairs. rest of the sac. It serves merely for the attachment of the short, thick, powerful muscles of the antennule which keep the latter in almost con- stant motion, and has probably nothing whatever to do with the censie functions of the otocyst. | b. Sensory Cushions, Of the three projections noted, the remaini two are sensory and bear sensory hairs (Plate 10, Fig. 55, set. set. fil.). The smaller of these (set. ¢a.), located on the median porti PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 207 of the posterior wall of the sac, bears a number of hairs with hooked shafts. The surface bearing these lies in a nearly vertical plane. From its position and the shape of its hairs this prominence is comparable to the sensory cushions upon the surfaces of which the otoliths are lodged in Palzemonetes, Crangon, and the crayfish. Irregularly disposed matrix cells are situated in clusters immediately beneath the hooked hairs (Plate 10, Fig. 50), and deeper in the tissues are the ganglion cells of the nerve fibres which supply the bristles (Fig. 50, el. gn.). In the larval stages of the crab this sensory cushion is relatively much larger. It extends through half the length of the sac, and its hairs are in contact with the otoliths which the sac then contains. Pores of tegumental glands penetrate the chitin of this prominence, as they do that of the sensory cushions in the crayfish and lobster, although found in no other part of the sac. These glands secrete a substance which, in the larval crab, attaches the otoliths to the tips of the hairs. Their presence in the adult crab is evidence in favor of the homology of this cushion with that described for otocysts containing otoliths. The other sensory cushion is much larger, and is produced by a partial invagination of a portion of the median and anterior walls of the sac, which forms an oval prominence (Fig. A; Plate 9, Fig. 48; Plate 10, Fig. 55, set. fil.). It is nearly 0.5 mm. in diameter, and its surface, making an angle of about 45 degrees with both the transverse and sagittal planes of the animal, inclines backward, inward and down- ward (Plate 9, Fig. 45). Its ventral portion is shown in transverse section in Figs. 47 48. The chitin of this cushion is very thin ; upon it is arow of long delicate hairs, called by Hensen (’63) “ Fadenhaare,” or thread-hairs. This row runs down somewhat obliquely from the upper side of the prominence to its ventral margin near the floor of the sac, its dorsal end being the more anterior of the two (Fig. A, set. jil.). This sensory cushion is also found in the sac of the larva, and the bristles it then bears are similar to those found projecting free into the lumen of the lobster otocyst from its median wall (Plate 5, Fig. 26, set. m.). The prominence we are now describing in Carcinus is probably therefore simply a further differentiation of the slight projec- tion noted in the sac of the lobster. Matrix cells send delicate processes into the hairs, as in those of pre- ceding species; the ganglion cells are situated directly beneath the hypodermis, but some distance posterior to the bases of the hairs (Plate 10, Fig. 53, el. gn.). No gland pores are present, nor are they needed, as the thread 208 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. hairs are never in contact with the otoliths, even in the larval stages. A third region, on which sensory hairs are located, is found at the ; extreme lateral side of the sac, beneath the fused lips of its opening (Fig. 4; Plate 9, Fig. 42, set.’). There is only a slight prominence, the surface bearing the hairs being nearly flat. The hairs are arranged in irregular fashion, somewhat like the groups of otocyst bristles situated near the aperture of the sac in the crayfish and lobster. Numerous a groups of matrix cells lie directly below these hairs, but no nervous” structures could be distinguished in their vicinity. The great hammer-like prominence, which serves for the attachmel of the antennular muscles, separates the sac roughly into an upper. anterior chamber and a lower, posterior one. The first of these com- — partments is again partially separated into two by the anterior sensory _ prominence, which nearly meets the “hammer.” These three chambers, — into each of which sensory hairs project, were likened by Hensen to the 4 semi-circular canals of the vertebrate ear, and the sensory regions to the criste acustice. As the compartments are in free communication, are— not at all canal-like in form, and are arranged in no definite positions relative to each other which might be of functional importance, there ] seems to be no more logical reason for making such a comparison than for comparing the hammer-like projection of the otocyst to the malleus, — The apparent division of the otocyst into three compartments is not a modification for the purpose of increasing its usefulness as a sense organ, but evidently a condition brought about mechanically by the differe tiation of the “hammer” along lines which would make it better adapted for the attachment of muscles. c. Structure of Hairs. The hairs, as already indicated in deseribi the sensory regions, are of three kinds. Hensen’s account of them is fairly good. He divides them into the following classes: (1) hook hairs (Hakenhaare), (2) thread hairs (Fadenhaare), and (3) rou hairs (Gruppenhaare). (1) The hook hairs are found on the posterior vertical cushion (Fig. A and Plate 10, Figs. 50, 55, set. ¢a.) arranged in a very irregular curved row. They vary from 25 to 31 in number, and are relatively very small, averaging 49 u in length and 4. in diameter. Their shafts are hooked, often bent nearly double, and are sparsely fringed near the tip, if at all. The base is enlarged, as is usual in otocyst hairs, but not so markedly as in the forms already studied. Instead of being attached to a large spherical membrane, the base of the shaft is set | PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 209 into a cup-shaped depression and so labilely fastened to the chitin of the sac wall (Plate 10, Fig. 51) that the hair can sway freely in any direction, as if it were attached by a ball-and-socket joint. This cup- like depression is characteristic of all the otocyst hairs of Brachyura. The hook hairs are present in the otocyst of the Megalops larva of Carcinus, and are there relatively much larger ; they extend over a large portion of the posterior end and floor of the sac, the curved row of 25 to 30 hairs occupying two-thirds of its length. As the otocyst is open at this stage, it contains numerous otoliths, and these are either im contact with, or attached to, the tips of these hairs. Measurements of a number of these larval hairs were made in the Megalops and the stage succeeding it, and a comparison of these with the same hairs of adults is made in the following table : . Average Length Average Diameter bees ot of ten Hook of ten Hook vets Hairs. Hairs. Adult (30 cm. long) 1.96 mm. 49 uw Young crab 0.24 mm. 47 Megalops larva 0.21 mm. 46 u This table brings out the interesting fact that the hook hairs of a Megalops larva, of a young crab and of an adult are of nearly equal size, although the otocyst of the adult is nearly ten times as long as that of the Megalops, and over eight times that of the young crab. On measuring the thread hairs to see if the conditions there were the same, it was found that in the adult they were three and a half times as long as in the Megalops stage; the thread hairs thus more than tripled their length, while the hook hairs remained constant. The number of hook hairs is approximately the same in the Megalops otocyst and in the sac of the adult. Their arrested development may be explained by the fact that they are true otolith hairs ; when the otocyst becomes permanently closed, otoliths can no longer enter the sac, and these hairs, as they lose their original function, do not grow pari passu with the other hairs of the otocyst, but remain unchanged. They do not degenerate and become entirely functionless, for they are still innervated in the adult crab, and, though sac after sac is shed and new ones formed without an otolith’s finding its way into the organ, they still retain the peculiar form of the original otolith bristles. 210 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. We are thus led to regard the hook hairs of the crab as homologues of the otolith hairs of Macrura, and for these five reasons:— (1) The similarity in their structure. (2) Their similarity in position at the posterior end of the sac. (3) Otoliths are in contact with the hook hairs in larval stages, though not in the adult. (4) When the otoliths — disappear, the development of the hook hairs is arrested. (5) Gland — pores open through the chitin of their cushion, as they do through that — of the crayfish and lobster, although they are not found in the other sensory regions of the sac. ‘ (2) The thread hairs are the largest, the most highly differentiated, and probably the most active sensory bristles of the otocyst. There are about thirty of them, arranged upon the large anterior sensory cushion in a regular row (Fig. A, set. jil.). These hairs are extremely attenuate. Measuring only two or three u at the base, the straight or slightly bending shaft averages 320 in length; it is unfringed save at the very tip, where for a short distance it bears two rows of ex- tremely delicate pinnules. A peculiarity of this fringed tip is that it is not a continuation of the main shaft of the hair, but seemingly a diminutive hair in itself, sprouting from the latter. It makes a slight angle with the main shaft, the end of which projects a short distance beyond the base of the offshoot (Plate 10, Figs. 53, 54). | The shafts of these hairs are directed out laterally, and slightly pos- teriorly, into the fluid contents of the sac, and they are so delicately attached at their bases that the slightest jar imparted to the liquid in which they float is sufficient to set them swaying. In alcoholi¢e material they break off very easily. The shaft decreases somewhat in diameter towards its base and then suddenly enlarges. This enlarge- ment is attached to the floor of a deep cup-like socket, the orifice of which is large enough to give ample play to the shaft in its movements (Fig. 53). ; Straight attenuate hairs are found in the otocyst of the Megalops larva having the same relative position in the sac as the thread hairs of the adult. These hairs are aot in contact with otoliths, but each shaft is fringed with filaments throughout its whole length. They become differentiated in later stages into the peculiarly modified thread hairs. Hairs similar to those of the Megalops larva just described are also found in the otocyst of the adult lobster, situated on the median wall of the sac and projecting free into its lumen. They are similar in both larva and adult, and are probably in function accessory © to the otolith hairs. They may be homologues of the thread hairs, | ~ PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 211 which, in the crab, with the disappearance of the otoliths, have taken on the chief functional activity of the otocyst, formerly vested in the hook hairs. (3) The group hairs (set.') form the third and most numerous class of the otocyst bristles of Carcinus. Irregularly distributed in the most lateral corner of the sac (Fig. A,) on a flattened portion of the wall ventral to the closed margins of the aperture (Plate 9, Figs. 42, 47; Plate 10, Fig. 55), they are unlike any of the otocyst hairs found in Macrura, being short, thick, and blunt, without a trace of fringing fila- ments (Plate 10, Fig. 49). They are 110 to 135 long and 124 to 14uin diameter. There are nearly 200 of these hairs, forming one large irregular group. They do not occur in the Megalops otocyst, therefore they must be developed at some later period. They may possibly be degenerated tactile hairs which in the formation of the otocyst have been folded into its cavity. Their proximity to the aper- ture of the otocyst makes this supposition highly probable. Their shafts are set into depressions in the sac wall, and, like the other oto- cyst hairs, they can sway freely on their bases. d. Formation of Hairs. The hairs are formed in Carcinus, and in the Brachyura generally, after the method already described in Palemonetes. From the presence of a cup-like depression at the base of each shaft, instead of the large spherical membrane found in the Macrura, it might be inferred that the cup results from the in- complete evagination of the hair. e. Otoliths are entirely wanting in the adult otocyst, but are present in those larval stages where the sac is still open. They consist, as usual, of grains of sand, which in this case are very small, for the sac itself in these stages is less than 0.3 mm. in length. They can readily be introduced into the otocyst of the Megalops, as its aperture is rela- tively large. When in a succeeding stage the sac is cast off with its otoliths at ecdysis, the aperture of the new cyst closes at once, and no foreign particles can enter it ; henceforth it is without otoliths, 2. Innervation of the Otocyst. The general course of the otocyst nerve is shown in Plate 10, Figure 55 (n. ot.). As in the forms previously described, the sac lies in close proximity to the brain, and its nerve is consequently short. It is given off with the antennular nerve from the anterior end of the cen- tral organ, and its course for a short distance is directly lateral, until the base of the antennule is reached. At this point the antennular eA BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. nerve (z. a@f.1) turns straight forward, while that of the otocyst divides — into three branches (Fig. 53, n. of., n. ot.!, n. of.!). The most median ~ and largest of these runs forward to supply the thread hairs; the — middle branch goes directly to the posterior sensory cushion, which — bears the hook hairs; while the third and lateral offshoot takes a nearly straight course along the posterior wall of the sac and supplies — the tactile hairs of the antennule, and possibly the group hairs of the otocyst. The ganglion cells of the hook hairs are some distance pos- — terior to the hairs and arranged in an irregular scattering group { (Plate 10, Fig. 50, cl. gn.). Those of the thread hairs are lateral and N posterior with reference to their hairs, lying immediately beneath the hypodermal cells of the sensory cushion, and forming an irregular single — row, which is nearly parallel to the row of thread hairs (Plate 10, — Fig. 53, el. gn.). a. Number of Nerve Elements to a Single Bristle. The nerve ele- ments of the thread hairs were brought out clearly and completely by methylen blue and by Vom Rath’s platinic-chloride method. The conditions found in a number of preparations are shown in Figure 53, where there is but a single element for each hair. This particular preparation was obtained with methylen blue, but the results were verified by Vom Rath’s method. Counted in serial sections, the num- ber of hairs and ganglion cells were approximately equal. : By the same method of counting, the elements of the hook hairs gave like results. In one case there were thirty hairs and thirty- one cells. No ganglion cells could be made out near the group hairs, nor any fibres supplying them. Certain clusters of cells are found directly beneath their bases, but their large peripheral processes, irregu= lar outlines, and lack of central fibres marked these as matrix rather than nerve cells. Here in Carcinus, then, as in the macruran forms described, there is but one nerve element to each otocyst hair. The distal segment of the antennule was by chance sectioned in making preparations of the otocyst, and when stained with iron hema toxylin, the innervation of the olfactory hairs found in that region was sharply brought out (Plate 10, Fig. 52). As in the examples of this _ type of hair already described, a large spindle-shaped group of about 100 — ganglion cells sends a strand of nerve fibres to the base of each shaft, These cells are relatively small and situated 0.5 mm. posterior to the hairs they supply. In Figure 52 a single nerve element is shown | diagrammatically in black. a ~ ie > y “4 PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. Al NS b. Peripheral Terminations. As seen in Figure 53 (Plate 10), the terminal fibres going to the thread hairs enter the pore at the base of the cup-shaped depression, pass up into the enlargement of the hair shaft, and there end free. In fact, there is in these hairs no functional neces- sity for the further continuance of the fibre into the shaft. Since the hairs project free into the liquid of the sac, if the otucyst is jarred or tilted, the shaft does not itself bend, but sways backward and forward upon its base. It is therefore at the base that the stimulus must mani- fest itself, and it was there in every case that the tibres were found to end. In the olfactory hairs, on the other hand, the nerve fibres continue up into the large hollow shafts for some distance (Plate 10, Fig. 52, set. o/f.). The olfactory hairs of Carcinus thus differ in their innervation from those of the otocyst, both in the number of nerve elements supplying each hair, and in the peripheral nerve endings. In the bristles of the otocyst there is but a single nerve element, and it ends free at the base of the hair without branching. In the olfactory hairs there may be a hundred ele- ments or more which end in the shaft of a single hair. e. Central Terminations. Entering the brain in front of, and just median to, the globulus, and ventral to the optic centres, the fibres of the otocyst nerve run straight back and enter the fibrillar mass (Plate 10, Fig. 55, n’pil. at.1), called ‘the neuropil of the first antenna” by Bethe (97), who has described the central endings of the antennular nerve of Carcinus. ‘The fibres of the antennular nerve end in a connected neuropil just median to those of the otocyst. Bethe judged from his physiological experiments that there should be certain fibres from the otocyst ending in the globulus. He was not able to demonstrate such endings with methylen blue, nor was there any evidence of their exist- ence in my preparations. According to Bethe the fibres from the oto- eyst end by the separation of their fibrillz in the neuropil. Lack of fresh Carcinas material prevented the verification of his work, but I have described similar conditions in the shrimp and crayfish. d. Histology of the Nerve Elements. As the finer structure of the elements of the central nervous system has been fully described by Bethe (98), it is unnecessary for me to say anything on that matter, and only a few words need be added here as to the histology of the peripheral nerves and cells. The peripheral nerve fibres are much smaller than in Palemonetes or Crangon, and are without a myelin sheath. The peripheral ganglion cells are relatively large, averaging 12 # in diameter. They are of the typical bipolar form, and are much elongated (Plate VOL. XXXVI. — NO. 7 4 214 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 10, Fig. 50, el. gn.). Their nuclei are nearly spherical, and contain at least one large deeply staining nucleolus. No special preparations were made for the purpose of demonstrating fibrillz in either nerve cells or” fibres. Bethe found them in all fibres, and traces of them in the cells of - the brain. \ 3. Development of the Otocyst. For the purpose of comparison with development in the lobster, the antennules of the first five free swimming larval stages of Carcinus were dissected out, stained and examined zn toto. By this means it was ascertained that there is no functional otocyst in the Zoea stages. (a) The first Zoea shows no trace of invagination in its antennule. There is, however, an aggregation of nuclei beneath the chitin of the region where the otocyst is to appear. (6) The second Zoea shows a slight depression on the dorsal side of the antennule, and its basal portion has begun to widen. (c) In the third Zoea this widening has increased, and the lateral wall of the antennule has now formed a rounded protuberance. The invagi- nation has increased in size and depth, but no hairs nor otoliths are yet contained in it. (d) At the Megalops stage we find that a sudden development has’ taken place, as in the fourth larval stage of the lobster. The Zoea has by a single moult become metamorphosed into a Megalops, and the oto- cyst changed from a shallow depression to a nearly closed sac, contain= ing sensory hairs and otoliths. Two sensory cushions are present: one e of these, posterior and median, bears 25 to 30 hooked hairs, upon the tips of which otoliths rest ; the other prominence projects from the anterior portion of the sedan wall, and bears a vertical row of about 30— hairs, the shafts of which are directed laterally. These hairs are long, attenuate, and well fringed with delicate filaments. They do not come into contact with the otoliths, and, as already noted, they develop into the thread hairs of the adult ; those of the first sensory cushion described correspond to the hook hairs of the mature crab. The third type of hair found in the adult is not developed at this stage. The aperture is anterior and lateral in position, and extends transversely across the antennule. (e) The next stage examined was that of a young crab probably of the stage immediately succeeding the Megalops larva. The otocyst is slightly larger, and its opening is already nearly closed. As a result, only a few small otoliths were contained in it. PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 245 The otocyst of Carcinus thus resembles very closely in its development that of the lobster. In both there is no trace of the organ in the newly hatched larve, and for three successive moults it is not functional. In the fourth larval stage, with a sudden metamorphosis of the animal’s general form, the otocyst is also rapidly changed from a mere depression to an active, well-developed organ. The significance of these sudden correlated transitions will be seen when the otocyst is considered physiologically. C. THEORETICAL CONSIDERATIONS. 1. Comparison of the Otocyst with the Vertebrate Har. The otocyst has been compared by many investigators to the auditory organ of vertebrates. Leaving their functions entirely out of account, how far do the two correspond in structure ? The otocyst of Macrura consists of an open sac, a sensory prominence, bristles, and otoliths resting upon them ; essentially the same conditions as are found in the ear of Myxine, though the latter has five sensory regions instead of one. The otocyst of macruran decapods might thus be well compared to an isolated ampulla in the ear sac of Myxine, and the sensory cushion to a single crista acustica. In the Brachyura the organ is still more highly differentiated. The sac is closed, there are three sensory regions, and the hairs found on them project free into the lumen of the otocyst; otoliths are entirely wanting. The structure of the sensory apparatus is in this case similar to that of the cristz of higher vertebrates, and the sac itself resembles the utriculus. But there cs no portion of the decapod otocyst so differen- tiated as to bear more than a fancied resemblance to the semicircular canals, the middle ear, or the cochlea of higher vertebrates. Eagh crista acustica in vertebrates, however, is made up of separate elements, which may be compared to the sensory elements of the otocyst. Every auditory hair of the crista is developed from the exposed end of a Specialized epithelial sense cell, which itself forms the basal part of the hair, and is supported in position by the other cells of the epithelium. Tt has been shown by both Retzius (94) and Morrill (’98) that these epithelial sense cells of the cristze in vertebrates are not true nervous elements, as the auditory fibres are not continuous with them. Both the cell and its auditory hair taken together are to be compared to the bristles of the otocyst, in that they constitute a non-nervous end-organ. 216 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Their innervation is also essentially the same. In the vertebrate crista an auditory nerve fibre passing from the brain is connected with a bipolar nerve cell in the auditory ganglion, from whence its peripheral fibre ex- tends to one of the epithelial sense cells, ending with a slight enlarge- ment in close proximity to, or in contact with its base. The single fibre” supplying each end-organ is never directly connected with the cell, nor does — it ever run through it to the hair itself. The only difference between the - peripheral endings just described, and those of the otocyst, is that in the — hairs of the latter the fibres end free in the base of the hollow shaft, at_ the point where, from the structure of the hair, the greatest stimulus — would be produced ; while in the vertebrate end-organ the nerve process is applied to the convex under-surface of the basal cell, which would transmit stimuli with an equal degree of intensity to fibres in contact — with it at any point. The otoliths of the vertebrate ear are formed by secretion, while ‘ those of the crustacean otocyst are largely granules of sand taken into the sac from the exterior. In some Crustacea, however, such as the Mysidee, and in many other invertebrates, the otoliths are formed within ~ the sac. In all decapods the innervation of the otocyst hairs distinctly differs from that of the olfactory bristles, not only as to peripheral termina- tions, but also in the number of nerve elements supplying each hair. As has been previously noted, the stimulus is transmitted by specialized cells or hairs to the nerve fibres of both the otocyst and the vertebrate ear, and is never applied directly to their endings. In either case only one nerve element is usually in contact with the terminal sense cell, and this is apparently ample to carry the isolated nervous message to the brain. With the olfactory sense it is different; in both vertebrates and Crustacea the chemical stimuli which produce the olfactory sensations act directly upon the nerve cells or their terminal fibres. In vertebrates portions of the nerve cells are exposed at the surface of the olfactory epithelium. In crustacea peripheral fibres from the ganglion cells of the olfactory nerve end free in the hollow, perforate bristles. In Nereis and the earth-worm, Langdon (’95, ’00) has shown that the processes of the olfactory cells end free upon the surface of the cuticula, and com- pletely exposed to chemical stimuli; a similar condition has been shown by Lewis (98) to exist in two polychetous worms of the family Maldanide. ’ The large numbers of nerve elements ending in each olfactory tube or bristle of decapod Crustacea may be accounted for by the fact that : PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 217 the stimulating chemical substances occur as slight traces only. In order _ that a sensation may be perceptible, apparently a large number of olfac- tory elements must be stimulated at once, for the larger their number, the stronger should be the sensation produced. The olfactory bristles are located on the flagella of the antennules, a position most favorable for the reception of chemical stimuli, as the flagellum projects some distance in front of the animal and can be kept in constant motion. The number of the bristles is limited on account of the small surface to which they are necessarily confined, so that, if thousands of olfactory fibres are to function simultaneously, large numbers of them must be exposed to the chemical stimulus in the same hair. It is possible, too, that different nerve elements may be affected by different substances in solution ; and that consequently many olfactory elements are necessary for each hair, in order that different chemical stimuli may be perceived. 2. The Neuron Theory. The conditions found in the sensory nerve elements of the otocyst are favorable to the neuron theory, in so far as they confirm the generally accepted idea that the nerve fibres are each differentiated from a single nerve cell, and that fibre and cell taken together form a trophic unit. This conclusion is borne out not only by the structural conditions already described, where each fibre is connected with only one peripheral ganglion cell, but also by an experiment which I made by severing the otocyst nerve proximal to its ganglion; in this case after the lapse of a few weeks degeneration of the sensory fibres took place back into the brain. As to the modifications of the neuron theory recently proposed by Apathy (’97) and Bethe (’98),—that the neurons are connected by fibrille, — the fibrillar structure of the fibres is confirmed by my prepara- tions, though no fibrillee could be demonstrated in the nerve cells. In regard to the definite connection of the neurons with each other by con- tinuous fibrils, such as Apathy figures and describes in the Hirndinea, my preparations gave no positive evidence; but the fact that the cen- tral fibrillations of the nerve elements of the otocyst could not be traced to determinate endings, makes it quite possible that such a direct com- munication between motor and sensory neurons may exist. While Bethe proved that there were more fibrillz in a motor fibre than extended into its central ganglion cell, and also, that some fibrillz entered the fibre by one branch and at once passed out by another, in no case did he trace 218 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. a single fibril from one neuron into another. If such a connection between nerve elements had been demonstrated beyond a doubt, they might still be considered as distinct trophic units, and the interdigitating fibrils uniting them as the products of separate neuron cells. In the light of the important discoveries of Apathy and Bethe, however, the old view, that the nervous impulses are transmitted from sensory to motor aeurous by the simple contiguity of their dendritic processes, may have to be abandoned for the more reasonable assumption of direct fibrillar communication. PART II.— PHYSIOLOGY. As Bethe has well said, the best of anatomical knowledge concerning — an organ cannot be taken as certain evidence of its functions. It is — only after these functions have been experimentally demonstrated, that we may ascribe them with confidence to the organ in question. Have we, then, any experimental proofs that the decapod Crustacea — hear? If so, is the otocyst the auditory organ ; if not, what is its fune- tion? These are the three chief questions which I shall attempt to answer. A. HISTORICAL SURVEY. Up to the time of Delage (’87) the auditory function of the otocyst was accepted, and that alone. Minasi (1775) promulgated the idea that Crustacea could hear. The hermit crab, Pagurus, was more sensitive than man to sound vibrations, The tones of a distant bell, the striking of a clock, were, according to this worthy monk, perceived by Pagurus sooner than they were by him. Aflianus (1784) notes that the fishermen of his time took Pagurus by means of music, All the older zodlogists have regarded the otocyst as an organ of audition. Hensen (’63) was the first to get experimental data. From the anatomical conditions found in the otocyst of the lobster, he argues as follows: Here are 468 auditory hairs upon which otoliths rest. Of these hairs no two are of the same size ; they vary in a nearly continu- ous series from 0.72 mm. to 0.14 mm. in length; thus the volume of the largest is to that of the smallest as 140: 1. Comparing these val PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 219 ratios to those of the volume of organ pipes, we should have, ¢f the hairs - responded to different sound vibrations, an auditory organ with a range of three octaves. To prove that his hypothesis was correct, sound waves were conducted, by a mechanical contrivance modelled after the middle ear of mammals, into the water of a vessel containing Mysis, the so-called auditory hairs of which were under observation by the microscope. When notes of a certain group were sounded on a musical instrument, a certain hair would vibrate and disappear from view. Others would also respond, but each to different sets of notes. Having proved that the different hairs responded to different sound waves, Hensen next determined that Crustacea would react to vibratory stimuli. A resonant bar of wood was floated in a vessel containing free- swimming individuals of the genera Mysisand Palemon. When the bar was struck, both forms responded by a strong leap away from the source of the sound. Palemon reacted even more strongly when rendered sensitive by gradual strychnine poisoning. Milne-Edwards (’76), Jourdain (’80), Delage (’87), and many others have accepted the sense of audition in Crustacea as a fact. Garbini (’80, p. 192) uncritically remarks: “Che i crostacei odano é indubitato ; lo sanno anche i pescatori, i quali devono avvicinarsi loro in silenzio” (That crustacea hear is undoubted ; this the fishermen know well, who, when they capture them, approach in silence). Individuals of Paleemonetes varians, which he kept in an aquarium, sprang backward at the slightest sound. Delage (87) was the first to discover another function than that of audition for the otocyst. By cutting off or destroying the sacs, he proved that they functioned also as organs of orientation. Animals so operated upon (Mysis, Palemon, and Polybius among Crustacea) were unable to keep their normal upright position in swimming. Blinding intensified the effect, showing that sight aided in orientation. The otocyst may therefore, in his opinion, be compared to the sim- plest form of the vertebrate ear, — that found in Myxime, — where the semicircular canals and utriculus serve the purpose of orientation, the sacculus that of audition (to intensity of sound). In the otocyst of Crustacea both functions are performed, he believes, by the same organ. Verworn (’91) proved that the otocyst of Ctenophores served simply for orientation, not being sensitive to sounds. Bunting (’93) confirms the conclusions of Delage as to the function 220 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. of the otocyst in geotropic orientation. When the otocysts of young crayfish were destroyed, especially if their chele were also removed to render their position in the water less stable, there was the same loss of power of orientation that had been observed by Delage. Kreidl (93), in order to avoid the disturbance to the normal condition caused by the removal of the otocysts, made use of the following in- genious experiment: Palemonetes newly moulted, and thus without otoliths, were placed in filtered water to which iron filings were added. The otocysts were soon filled with the metallic particles, the chele being used to convey them to the opening of the ear in the dorsal wall of the antennule. When now a strong electromagnet was held at one side of, and slightly above the sacs containing the iron otoliths, the shrimp would lean a little to one side, its dorso-ventral axis, normally coincident with the direction of gravity, pointing away from the magnet. This new position of the dorso-ventral axis is proved by mechanics to be the resultant of the two pulls, that of gravity and that of the magnet, the animal accommodating itself to the direction of the resultant of the two forces. If the magnet were held to the right of the animal, the otocysts would be stimulated in precisely the same way as by gravity alone when the shrimp’s dorso-ventral axis is artificially turned toward the right; the result is that it attempts to recover its normal position with reference to gravity, and thus turns its vertical axis away from the magnet. Kreidl, going a step further than his predecessors, affirms that the otocysts are not auditory, but exclusively static in function. Thus they should be called stato-cysts, not oto-cysts. Still further evidence as to their static function is supplied by Clark (96). The compensation movements of the eyestalks of the fiddler crab (Gelasimus pugilator) and the lady crab (Platyonichus ocellatus) were observed. Tilting a normal animal about its antero-posterior axis gave a parallel compensating movement of the eyes through an angle of 35° to 45°, whether the tilting was to the right or left. On rotation about the dorso-ventral axis, no such movements are shown, though when rotated about the lateral axis, the animal’s eyes moved in the opposite direction through an angle of 35°. If both otocysts were removed, these compensative movements were much reduced, and the general movements of the crab also became very uncertain. After removal of one otocyst 94 per cent of the animals showed on rotation toward the nuinjured side less compensation than uninjured animals. Blinding produced only a slight reduction in the compensatory a PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. ob motions, but when, in addition to this, both otocysts were destroyed, compensatory movements completely disappeared. Bethe (’97), in his physiological work on Carcinas mezenas, confirms Clark’s results. In a previous paper he (95%) observes that Mysis can hear after the otocysts have been destroyed, but with difficulty ; aiso that the animals are more sensitive to low tones than to high. Thus, until 1898 three views were held as to the function of the otocysts: (1) That they are purely auditory organs (Hensen and the earlier zodlogists). (2) That they are both auditory and static in function (Delage and Bethe). (3) That they are purely static in function, i. e. organs of orientation (Kreidl, Clark, and others). To determine whether decapod Crustacea really hear, and if so, whether the otocyst is the organ of audition, is the aim of two papers by Beer (98, ’99). In criticising the conclusions reached by Hensen and Bethe, Beer remarks in his first paper that, because decapods were made to react to different sounds, does not prove that these Crustacea responded to true sound, or that they heard. These reactions may have been due to their feeling vibrations transmitted to the water from the walls of the vessel in which they were confined, — a tactile reaction, or, to use Bethe’s term, a “tango-reflex.” Experiments with sounds produced in the air Beer considered superfluous, as it is a well-known physical fact that most of the sound waves are reflected from the surface of water. Beer found that Crustacea reacted strongly to sounds produced in the water by striking partially submerged bells, jars, etc., but only when they were not at a greater distance from the source of sound than that at which vibrations conld be detected by the hand immersed in water. The animals responded more strongly when near the walls of the vessel ; but vibrations could be felt by the hand also in this position more dis- tinctly, even though further removed from the source of the sound. For animals well supplied with tactile organs, he regards pure sound or pure audition as impossible; because vibrations could be felt as soon as heard, and, this being the case, audition would be useless. - On removal of the otocysts, Paleemon and Palemonetes still responded to sound waves produced in the water. There was, however, a slight in- hibition of the customary reactions, therefore the hairs of the otocyst are probably slightly tactile as well as static in function. 223 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. From experiments on many different species of Crustacea, Beer (’98, p- 31) concludes: ‘ Wir haben gute Griinde, dem in Rede stehenden — Sinnesorgane der Krebse statische Functionen zuzuschreiben, und haben vorlaufig gar keinen Anhaltspunkt, iam Horfunctionen, ja den Krebsen tiberhaupt GehOrsinn, zuzuschreiben.” Hensen’s statement that the free auditory hairs of Mysis vibrated to different musical notes is simply an interesting physical fact. Hairs on — the back of one’s hand will do the same, but they are not auditory, — The true sense of hearing is lacking not only in Crustacea, but probably in all other water-inhabiting animals lower than Amphibia, especially in — invertebrates. Beer thus comes back to the opinion of Johannes Miiller (’37) ex- pressed sixty years before: That in most invertebrates we find nothing comparable to the ear; and any reaction to sound vibrations should be attributed to a tactile rather than to an auditory sense. A few months later Beer (’99) brought out a second paper, describing his experiments with blind shrimps, and answering a criticism of his pre- vious work by Hensen (99). Here the auditory sense, he urges, ought to be intensified, all possibility of sight entering as a factor into the experiments being effectually eliminated. The conclusions reached by him in his earlier work are verified in this. General Criticism. It is a noteworthy fact, that in the experimental work done to deter- mine the function or functions of the otocyst, few of the investigators have acquainted themselves with the finer structure of the organ under consideration ; one of the essentials for successful physiological work is a complete knowledge of the anatomical side of the subject. This is well illustrated in Bethe’s work on the brain of Carcinas, where anatomical facts, obtained by means of methylen blue, laid the groundwork for his later confirmatory experiments. Since the dissections by Hensen, little or no morphological work has been done on the otocysts of the Brachyura, yet a deal of physiological work has been attempted. The experiments of Beer are beautifully worked out, and logical in sequence ; yet, while he tried experiments on water-inhabiting animals, no attempt was made to experiment on amphibious decapod Crustacea, such as the fiddler crab. These animals, spending, as they do, a good share of their life on land, would certainly have more need of an auditory organ than decapods which are always beneath the surface of the water PRENTISS: THE OTOCYSL OF DECAPOD CRUSTACEA. 223 B. EXPERIMENTS AND OBSERVATIONS. I. The Otocyst as an Auditory Organ. That the responses of water-inhabiting animals to atmospheric sounds is nothing more than a myth, has been too well proved by Beer to need further investigation. The well-known physical fact that the larger part of the sound waves are refiected from a liquid surface is enough in itself to confute fables of fishes and crustacea hearing, and coming to be fed at the sound of a bell. But since in the case of responses of decapod Crustacea to sound vibrations conducted into the water, the experiments of Beer contradict Hensen’s earlier results, repetition of Beer’s work, though perhaps not absolutely necessary, may not be out of place. METHODS. The shrimps to be experimented upon (Palemonetes) were placed in glass vessels 40 cm. in diameter and 20 cm. deep. Sound waves were conducted to the water by means of a steel pipe one inch in diameter and about two feet long, which was firmly clamped at its upper end and projected into the vessel containing the shrimps; a brass rod was in some cases substituted for the pipe. The pipe and rod were set into vibration either by striking them with a hammer, or by drawing across them, bowlike, a strap of rosined leather. Sounds were also produced by striking glass jars suspended in the water, and by striking the sides of the aquarium itself. The movements made in pro- ducing the sounds were completely screened from the view of the shrimps by pieces of cardboard placed over and at one side of the vessel, a small aperture being left for observing their reactions. Palzmonetes could be made very sensitive to all nervous stimuli by leaving them over-night in sea water containing from 0.1 to 0.2% of sulphate of strychnia. This solution is fatal to a small fish (Fundulus) in five minutes; many of the shrimps die, but the sensory apparatus of those which remain alive is rendered abnormally acute. Blinding was accomplished by simply painting the eyestalks with a thick coat of lampblack and shellac ; the otocysts were removed by means of a fine hooked needle, with scarcely any other injury to the animal. 1. Responses of Palcemonetes to Vibrations transmitted to Water. _@. Normal Conditions. Under normal conditions, when sound vibra- tions were transmitted to the water, normal animals responded by a 224 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. slight leap backwards or to one side, if.the source of the sound was within a distance of 20 cm. If an animal happened to be near the side of the vessel, and the sound was produced near the opposite wall 40 cm. distant, the response would be, not a durting away from the source of the sound, but a leap back from the side of the vessel toward the source of the sound. Again, if an animal was facing the side of the aquarium with its antennz in close proximity to it, and the opposite wall was sharply tapped with the finger-nail, or lightly with a hammer, the shrimp, as before, sprang away from the side of the vessel toward the source of the stimulus. The response was usually well marked, a leap of from 10 to 15 cm. being made. b. Povsoned with Strychnine. The responses obtained were invariably much stronger and more uniform with animals poisoned by strych- nine in the manner stated above, than with normal shrimps. In other respects they were the same, and served merely to emphasize the results obtained by the first experiments. Blinded individuals showed practi- cally the same reactions, but to make sure that the factor of vision was effectually cut out, the eyestalks of the shrimps in the succeed- ing experiments were all painted. c. Both Otocysts removed. Of animals from which both otocysts had been removed, all but one gave a more or less strong response to the sounds conducted into the water in which they were swimming. The reactions were not. as marked, nor could they be produced at as great a distance from the source of the sound, as in the case of normal animals. Nine individuals were affected by the stimulus when at a distance of about 10 cm.; the rest, only when in still closer proximity. A slight jar imparted to the walls of the aquarium produced essentially the same responses as the transmission of sound to the water by means of the vibrating pipe or rod. Removal of the otocysts has, therefore, only a very slight inhibitory effect upon the responses called forth by sound-wave stimuli in normal or strychnine-sensitized animals. d. Removal of Antenne and both Antennules. The removal of the antennee and antennules, which bear large numbers of delicate tactile hairs, very much reduced the reaction of the shrimps to these vibratory stimuli. Only when an animal was in close proximity (5 cm. or less) to the source of the sound, or in contact with the walls of the vessel; would it respond, and then only feebly. Slight jarring of the aquarium produced no reaction, unless some part of the animal’s body directly touched the sides or bottom of the jar, or was in contact with the sound- producing instrument. PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 225 The above experiments were duplicated on Crangon vulgaris with similar, though less marked results, as Crangon is much more sluggish than Palzemonetes. A third set of experiments was tried with Virbius zostericola, a shrimp-like decapod without otocysts. Normal animals responded vigor- ously on striking a glass jar partially submerged beneath the water in which they swam. This response, much increased by strychnine poisoning, was distinctly diminished when both antenne and antennules were removed. e. Meaning of these Experiments. All of my experiments confirmed the conclusion of Beer, that free-swimming decapods, whether possessing otocysts or not, will respond to stimuli which are transmitted to them by the liquid medium they inhabit. The next question is, to determine whether this response is caused by the perception of sownd waves or by the coarser vibrations or jars imparted to the water. In other words, have we to do with true audition or with the sense of touch? Beer has clearly shown that there is no such thing as the transmission of pure sound waves from air to water. Coarser waves are imparted to the liquid simultaneously with those of sound, and can readily be felt by the immersed hand. After making a number of trials with sounds produced as in the pre- ceding experiments, I ascertained that the vibrations not only could be plainly felt by the submerged hand, but also that they could be felt at a distance from 10 to 20 em. greater than that at which the shrimps would react. This fact does not at all prove that the animals experimented with do not hear, but merely shows that the responses suppvsedly pro- duced by sound stimuli may be simple tactile reflexes, called forth by vibrations which, since appreciable to the immersed fingers, we may cer- tainly assume to be felt by these animals, so well supplied with delicate tactile organs. That the reaction is really due to tactile stimulus rather than to audi- tion, is indicated by several facts brought out by the experiments: (1) Animals, when near the wall of the vessel, even though distant from the source of the sound, respond vigorously, leaping nway from the wall and toward the sound. The wall is set into vibration by the pro- duction of the sound, and it is apparently this vibration which affects them, rather than the true sound-waves imparted to the water. (2) The average distance from the source of the sound at which they will respond is less than that at which vibrations may be felt by the hand. i) 26 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. (3) Removal of the antenne and antennules which are supplied with numerous tactile bristles, inhibits the reaction. (4) Decapods, such as Virbius, normally without otocysts respond vigorously ; but removal of antennz and antennules diminishes their — sensibility in a marked degree. (5) Precisely the same responses as were called forth by the produc- _ tion of sound were also obtained by simply tapping or jarring the walls — ‘ of the aquarium. Whether due to tactile stimulus or to audition, the fact remains, that the otocyst has little or no part in producing the reactions observed in the series of experiments; for (1) decapods normally without otocysts respond as vigorously to the same stimuli as those possessing them, and (2) the removal of the sacs from the latter has only a very slight in- 3 hibiting effect, which might be due either to the loss of these organs, | or to the injury of the nerves supplying the many tactile bristles of the antennule. Consequently, the otocyst not being the organ by stimulation of which © responses to sound vibrations are called forth, and there being no other — sensory apparatus in Crustacea especially differentiated for the reception of sound waves, we are led to the conclusion that in decapod Crustacea — a true auditory organ is wanting. The acute tactile sense of decapods may to some extent serve the same purpose that audition does in vertebrates. In mammals the senses of touch and hearing grade into each other. The range of the average auditory organ in mammals is from 30 to 16,000 vibrations per second ; waves of less than 30 vibrations per second do not usually produce audi- tory sensations, but are appreciable to the tactile sense. It is important to note that decapods respond most vigorously to low notes, and not at all to high notes or sounds produced by very rapid vibrations, This_ fact would seem to be good evidence that the vibrations imparted to the water and perceived by decapods correspond to those which produce tactile rather than auditory sensations in vertebrates. = 2. Responses of Gelasimus pugilator (Brachyuran decapod), a. To Vibrations transmitted to Water. On the conduction of sound waves to water by the same means as in the preceding experiments, these fiddler crabs responded, but by no means as vigorously as did the Ma- crura. They always rested upon the bottom of the aquarium, and reacted by retiring slowly, either from the source of the sound, or from the vibrating walls of the aquarium. In either case the response took PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. ei f place only when the animal was within a few centimetres of the vibrat- ing surface, and was most marked when the antenne and antennules were in close proximity to it. After blinding the animals and removing their otocysts, no apparent difference could be detected in the reactions called forth, as compared with those of normal crabs; removal of the first two pairs of appendages caused, on the contrary, the responses to almost completely disappear. b. To Atmospheric Sounds. As the fiddler crab is on land a large part of the time, a number of experiments were tried to determine the effect of aerial vibrations upon them when they were feeding under per- fectly normal conditions. A position for observation was selected near a bank which was completely honeycombed by their burrows, where one could see the animals perfectly well, and yet be screened from their view by intervening bushes. If one remained perfectly motionless, the animals would come within a short distance of the observer’s place of concealment, feeding as unconcernedly as if no one were near. Whena number of crabs were little more than five feet distant, a horn was blown, care being taken to direct it away from them. Although a sound was thus produced loud enough to be heard at some distance, all the animals continued to feed undisturbed. The striking together of two stones, and the sound produced by strik- ing an iron pipe with a stone (the objects in both cases being held in the hand) also had no effect upon them. On striking the ground with a heavy stone all the crabs within a radius of ten or twelve feet were startled ; some of them merely stopped feeding, while others scuttled into their burrows. The same result was brought about by simply stamping upon the ground. If a quick movement was made in the sight of the animals, they at once scattered precipitately to their holes. These observations were repeated a number of times, and on crabs of two different localities, with the same results. From these experiments and observations, we may draw the conclusion that the fiddler crab, whether in water or on land, does not respond to true sound-stimuli, but is affected only by jars or vibrations transmitted to the water or to the ground. In neither case can they be said to hear. When feeding upon land they do not depend upon an anditory sense to protect them from terrestrial enemies, but rely entirely upon their keen vision and delicate tactile organs. The statement is generally accepted, that all animals which produce sounds also have a sense of hearing, and this is advanced as an argu- ment in favor of audition in Crustacea. The two well-known examples iw) 28 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. of sound production among decapods, observed by T. Parker (’78) and Goode (’78), are (1) the stridulation of the rock lobster, Palinurus, where the sound is produced by rubbing the second segment of the antenna against the antennule, and (2) the pistol-like report produced by Alpheus in snapping together the claws of the great chela. As Beer has pointed out, the otocyst is poorly developed in Palinurus ; furthermore, no in- dividuals of either species have ever been observed to respond in any way ve when these snappings or stridulations were produced. We can no more argue, from these two instances of sound production in decapods, that there is an auditory function in all Crustacea than we can that all fish hear because the drum-fish makes a sound. The enemies of water-inhabiting crustaceans produce no sounds which would reveal their presence to their prey; the latter would therefore have to rely upon other forms of stimulation for the detec- tion of their foes. Even if it were admitted that they possessed a sense of hearing, yet, as shown both by Beer’s experiments and by my own, it must be so restricted in range that they would be able to detect sound produced at no greater distance than that at which the vibra- tions could be felt by the hand. Such a dull sense as this would be of no practical value in protecting crustaceans from their foes. Both observation and experiment lead, then, to the following general conclusions : (1) The reactions formerly attributed to sound stimuli are nothing more than tactile reflexes. (2) The otocyst has little or no part in calling forth these reactions, (3) There is no direct evidence to prove that decapod Crustacea hear, and until such evidence has been obtained, we are not warranted in ascribing to the otocyst a true auditory function. II. The Otocyst as an Organ of Equilibration. All water-inhabiting, free-swimming animals which maintain a defi- nite position with reference to gravity either during locomotion or when at rest, can thus orient themselves only under one or the other of two conditions : Either the animal must be normally in a condition of stable equi- librium, keeping its definite position under the influence of gravity like any inanimate body; or, if a position of unstable equilibrium is main- tained, the animal must in some way be made sensible of the direction of gravity, and must keep itself in equilibrium by its own efforts. — y PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 229 In the first case merely the mechanical action of gravity is called into play; in the second instance, besides the outside action of a physical agent, a subjective sense of direction and orientation is involved. In free-swimming decapods the body, moving or at rest, is in a position of unstable equilibrium. The dorsal side being always kept uppermost, the centre of gravity is high up, and a dead individual or an inanimate object of the same size, form, and disposition of weight would at once turn over. These animals must then by some means be rendered sensible to the direction of gravity, in order to be able to maintain a definite position of unstable equilibrium with reference to it. To determine what are the organs which perform the function of equilibration, the following means have been employed in the present investigation : (1) Removal, or prevention of the action of an organ, and observa- tion of the effects on the equilibration of swimming or walking decapods. (2) Observation of the effect of such removal on the gimbol-like movements of the eyestalks (compensation movements) when the animal is rotated about its different axes. (3) Observations on the orientation of animals normally without otocysts. (4) The effect of the development of the otocyst on the equilibration of the free-swimming larve. (9) The effect on equilibration of the addition of magnetic attraction acting on the otocyst at right angles to the pull of gravity. In these experiments blinding was accomplished by painting the eyestalks with a mixture of lampblack and shellac. The otocysts were removed under the lens of a dissecting microscope with the aid of a fine needle, bent in the form of a hook. Other parts, such as flagella of antenne and antennules, were simply cut off with a pair of fine scissors. Palemonetes vulgaris, being hardy, was the species chiefly employed, but experiments of a like nature were also carried on with Mysis, Crangon, and Gelasimus. A large number of trials were made with each species. When organs were cut off or destroyed, the animals so operated upon were kept under observation for from 15 to 25 days, and the experiments were then repeated, in order to make Sure that the effects observed directly after the operation were not due to abnormal conditions produced by nervous shock. VOL. XXXVI. — NO. 7 5 230 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 1. The Removal of Sense Organs and its Effect on Equilibration. The normal position in which a shrimp, like Palemonetes, holds itself while swimming, is very characteristic : (a) The dorsal side of the body is always kept uppermost, its dorso- ventral axis corresponding to the direction of gravity, and its long axis usually lying in a horizontal plane. (6) Shrimps can be overturned only with difficulty, and even if this is accomplished, they right themselves at once. (ec) Animals coming to rest upon surfaces not horizontal tend to keep themselves in the horizontal plane, but with the dorsal side always up. : a. Eyes blinded. Nearly fifty animals were operated upon in this way and their movements observed. Placed in an aquarium, they swam about indiscriminately, but always with the dorsal side up, there being little if any rolling from side to side. They were not easily overturned artificially, and when interfered with, righted themselves quickly. The most noticeable difference to be observed between their movements and those of normal animals was the tendency to remain quiet and to hold fast to any object with which they came into contact, thus substituting the sense of touch for that of vision lost. It is apparent, therefore, that some organ or organs other than the eyes play the chief part in equilibration. b. Both Otocysts removed. Twenty-five animals were operated upon by removing both otocysts. In swimming there was still a strong tendency to keep the dorsal side uppermost, but there was in every case marked rolling from side to side, which occasionally culminated in a complete rotation about the long axis of the body. The animals could be easily overturned, and though they strove to right themselves, it was not accomplished as soon nor as accurately as in normal or blinded shrimps. They were more apt to remain quiet, or to swim along upon the bottom of the aquarium, than to swim free. If the long flagella of the first and second antennz were removed, rolling motions were increased and also the difficulty in righting themselves if overturned, the flagella being probably used as balancing organs in equilibration ; but the extirpation of the otocysts alone brings about a marked loss of orientation, much more pronounced than that produced by simply blinding. c. Both Eyes blinded and both Otocysts removed. Upon removal of both otocysts and blinding of both eyes, entire loss of the normal PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. Zoe position in swimming resulted in twenty-one trials out of the twenty-five made. The animals turn over and over, rotating about the long axis, now in one direction, now in the other; they also pitch forward and backward about their transverse horizontal axis, and often swim upon their backs. They do not resist overturning, unless holding to some stationary object, and make no attempt to right themselves when swim- ming free. The moment they come in contact with a horizontal sur- face, such as the bottom of an aquarium, they at once take up their normal position, righting themselves quickly, but if the surface they touch be oblique or vertical, and even if they come in contact with the under side of a horizontal surface, they cling to it tenaciously, taking up a position with reference to the plane of contact, and not in relation to the direction of gravity, as is the case with normal ani- mals. Thus the phenomena of orientation completely disappear in the majority of cases when both otocysts and eyes are rendered func- tionless, at least in the free-swimming animal. When the animal comes in contact with solid objects, the sense of touch asserts itself and the phenomena of orientation are again, to a certain degree, made manifest. d. One Eye blinded, both Otocysts removed. The conditions here are essentially the same as when only the otocysts are extirpated. There is a well-defined rolling motion in swimming, and if overturned artificially, the animal is very slow in regaining the original position. e. Both Eyes blinded, one Otocyst removed. In such experiments no effect was produced different from that brought about by blinding alone. There was no evidence of a tilting of the dorso-ventral axis toward the injured side, as might be expected, if the functions of the two otocysts were co-ordinated. Nor was there during swimming a rotation toward the side from which the otocyst had been re- moved. We may therefore conclude that in the phenomena of equi- libration each otocyst, as well as each eye, acts independently. As check experiments, both antennules were removed distal to the otecysts. No abnormal conditions were produced in swimming move- ments, the wounds healed, and these individuals lived in aquaria as long as normal animals. Where the otocysts were extirpated, individ- uals were kept as long as four weeks, and after this interval, when blinded, they gave the same evidences of loss of orientation as they did immediately after the operation. These observations, made upon Palemonetes, were found to hold true also for Crangon, Mysis, and lobster larve. Experimentation with the 232 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. fiddler crab gave like results. If blinded and deprived of otocysts, the erabs rolled both forward and backward when walking or running; this effect was still more apparent when the animals were placed in the water. 2. Removal of Sense Organs and its Effect on the Compensation Movements of the Eyes. The following experiments, carried out on Gelasinus pugilator, confirm the work done by Clark (96). When a crab is tilted to the right or left, forward or backward, the eyestalks tend to keep their original direc- tions, thus seemingly moving through a certain angle. Such move- ments, which have been observed also for the head and eyes of many vertebrates and insects, are called compensation movements, and the angle of movement, the angle of compensation. The angle of compensation in the fiddler crab was measured by means of the apparatus described by Clark (’96), a small table to which the animals could be securely fastened and tilted about their chief horizontal — axis. A scale ruled to degrees enabled one to read accurately the angle of compensation, and the angle through which the animal was turned. The long eyestalks of the fiddler crab make it easy to determine the angle of the eye movements. The angle through which the animals were turned was in all cases 45° first to the right, then to the left, about the chief, or longitu- dinal axis of the body. In each experiment fifty animals were used, the average being taken as the angle of compensation. These animals were most of them kept twenty days, and the angle then measured again, thus guarding against abnormal conditions. a. Normal Animals. In normal crabs the eyestalks are so held as to make an angle of about 22° with the vertical. The eye movements are always correlated, and if the animal’s body is tilted to the right (45°) the right eye makes a compensating movement of 18° upward, the left eye one of 25° upward; rotated to the left, the conditions are just re- versed, the right eye now moving through an angle of 25°. The move- ment of the eye of the side toward which the animal is rotated is in each case less by about 7° than that of the other eyestalk. This is due to interference of the carapace with the eyestalk, preventing its passage through a greater angle. b. Both Eyes blinded. Tilting either to right or left had the same general effect as in normal animals, but the right eye described an are of only 13°, the left eye one of 20°, or vice versa. There is thus a PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 233 marked reduction in the angle of compensation, a decrease of about 5°, as compared with normal animals. This shows clearly the extent to which vision enters into the orientation of these animals. e. Both Otocysts removed. The angle of compensation is here reduced to 3° and 5°, respectively, for the eyestalks on the side toward and from which the rotation takes place. Even without rotation the positions in which the eyes are held are not definite, as they are in animals which possess otocysts. The stalks often make an angle of 40° or more with the vertical, and their movements are no longer correlated. This, together with the marked decrease in the angle of compensation, as compared with that of blinded animals, makes it evident that in equilibration and orientation the otocyst plays a much more impor- tant part than does the organ of vision. d. Both Eyes blinded and both Otocysts removed. Onrotation it was found that the compensatory movements of the eyestalks were practically wanting. Two individuals only out of fifty showed movements of from 3° to 5°. In the greater number of cases no movement could be de- tected, and in the remainder the angle averaged lessthan 1°. There was a still greater tendency for the eyestalks to be held in indefinite positions when at rest, and at unequal angles. Fifteen such individuals were kept in an aquarium more than twenty days, and after this lapse of time practically the same results were obtained, showing that the shock of the operation of removing the otocyst had no effect upon the results of the experiments. Furthermore, removal of the antennules distal to the oto- cysts had absolutely no inhibiting effect upon the movements of the eyestalks. This series of experiments corroborates, as far as they go, the conclu- sions of Clark (’96). It is clear from them that the otocyst is the chief organ in equilibration, though sight also plays an important part in the orientation of these animals. Since the above work was done (July, 1899) a paper has been pub- lished by Lyon (99) on the comparative physiology of compensatory movements. These movements were studied by him in many vertebrate and invertebrate forms; they were found to exist in insects as well as Crustacea. Using the crayfish, he confirms Clark’s results to some extent, but finds that on blinding the animals and removing the otocysts a con- siderable angle of compensation still persists. This, together with the fact that insects, which lack otocysts, show the characteristic movements, he uses as an argument against the otocyst being an organ of equilibra- tion. Lyon also finds that upon rotation about a vertical axis there isa 234 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. compensation movement of the eyestalks of the normal crayfish through an angle of 10° to 18° ; and, further, that when the animal is rotated about its long axis blinding causes a diminution of 10% in the angle of compensation. His results therefore give a much more important place to vision in orientation than do the conclusions of Clark and myself. However, from the combined results of the experiments of Clark, Lyon, and myself, one cannot avoid the conclusion that, in the fiddler crab at least, the otocyst is by far the most important organ in equilibration ; next in order comes vision, and then muscular and tactile sense. 3. Lquilibration of Animals normally without Otocysts. Virbius zostericola, a shrimp quite common at Wood’s Hole, Mass., does not possess otocysts. Observation and experiment brought out several interesting facts concerning it. In the first place, it is not a free- swimming form. Its normal habitat is on the eel grass, to which it clings in positions indifferent to the direction of gravity. When forced to swim, it does so in a very uncertain manner, with the dorsal side usu- ally uppermost, though this is a position of unstable equilibrium, If overturned artificially (and this is easily accomplished), it rights itself slowly and will cling to the first object it may chance to touch, Re- moved from its supporting blades of eel-grass, its unstable manner of swimming closely resembles that of shrimps in which the otocysts have been destroyed. If the eyestalks are painted with lampblack, and the animals so treated are placed in a large aquarium, and forced to swim, apparently all sense of direction and means of orientation are lost. 4. The Effect of the Development of the Otocyst on the Equilibration of Lobster Larve. As has been shown in the morphological part of this paper, there is no otocyst in the newly hatched larva of either Palemonetes, the lobster, or the crab, nor is there a functional organ during the first three larval stages. It begins to invaginate only in the second larval stage, and it is merely a shallow cup-like depression in the third stage; not until the next moult do the sensory hairs and otoliths appear. When we examine the conditions as to equilibration and manner of swimming in the different larval stages, we find that in the first larva the body is not definitely oriented while swimming. Newly hatched lobsters are very unstable in their movements, often swim or come to rest upon their backs or sides, and show a tendency to roll from side to side while swimming. The animal swims by means of the exopods of the ee ees ip Ral EE Syn —- PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 235 thoracic appendages; the abdominal segments are flexed ventrally, and the thoracic endopods, hanging down, steady the rolling motions some- what. In the second stage the conditions are essentially the same. In the third stage the larve are more stable, though the otocyst is still functionless. This greater stability is explained when the Figure B. Lateral view of lobster larva of the third stage, showing swimming position. Magnified 6 diameters. Swimming position of the body and appendages is observed (Fig. /). The thoracic appendages are now relatively large, as compared with the size of the body. They are allowed to hang down ventrally, and in conjunction with the curved condition of the abdominal segments, serve to lower the position of the centre of gravity in the whole animal, thus rendering its swimming position much more stable. A= = Ficure C. Lateral view of lobster larva of fourth stage, illustrating the change in swimming position due to the presence of a functional otocyst. Magnified 6 diameters. Turning now to the fourth larval stage, we find the swimming posi- tion of the body entirely changed (Fig. C). The abdomen is no longer flexed and curved ventrally, but is held in approximately the same horizontal plane as the cephalothorax, while the thoracic appendages, instead of dragging downward through the water, are held up and for- ward in a line parallel with the long axis of the body. The great chelz project in front like the arms of a person preparing to dive, the exopods 236 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. of the thoracic appendages have been lost, and the larve now swim swiftly by means of the abdominal swimmerets. Although, from the position in which the body and appendages are held, the larva is in unstable equilibrium, it now orients itself very definitely during locomotion, in sharp contrast to the preceding stages. All signs of rolling from side to side, or pitching forwards, are com- pletely lost. The larva swim straight ahead with the body held usually in a horizontal plane and dorsal side up. The same position is also invariably maintained when the animals come to rest. Thus this sudden change as to form and swimming position in the fourth larva, unfavorable though it is for equilibration, is yet accom- panied by more delicate powers of orientation, and greater stability in swimming than are met with in the three earlier stages, where the centre of gravity of the animal is lower. Bearing in mind the fact that the otocyst first becomes functional in the fourth larval stage, we can only conclude that an intimate connection exists between its appearance as an active organ, and the delicate static sense which is suddenly exhibited by the larvze. If larve of the first, second, and third stages are blinded, their powers of orientation are almost entirely lost, but the same experiment has little or no effect upon the equilibration of the fourth larva. The first three stages thus depend mainly on vision for their imperfect ori- entation ; in the next stage this function has been largely transferred to the otocyst. A similar correlation between the development of the otocyst and the appearance of a static sense is found in the metamorphosis of the crab. The pelagic unstable Zoea larva is without otocysts, while the Megalops larva, which exhibits perfect powers of equilibration, possesses these organs well developed, and even containing otoliths, which are absent in the sac of the adult. The correlation which evidently exists between the formation of the functional otocyst and the sudden increase in static powers exhibited by lobster larvee is particularly well shown in the marked alteration in the swimming position maintained by the fourth larva, as compared with that of the three earlier stages. Previous to the fourth stage, the lack of a delicate static organ is compensated for by the maintenance of an attitude in swimming which increases the stability of the moving body. Just as Bethe (’95) found that Mysis, deprived of its otocysts, would after an interval of some days recover its power of orientation by curving the abdomen upward and thus. by lowering the centre of gravity, PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. Za put the body in natural equilibrium, so in the case of the first three lobster larve, the attitude maintained is an adaptation for the greater stability of the free-swimming animal, as yet without static organs. But when, in the next stage, the otocyst becomes functional, such an adaptation is no longer necessary, and the sudden change to the un- stable swimming position of the fourth larval stage results (Fig. C). This is the more natural attitude, and is advantageous to the animal in that it allows of greater speed in swimming. 5. The Function of the Otoliths. At the time when the otocyst was regarded as an auditory organ, the otoliths were supposed to act simply as intensifiers of the sound vibra- tions, but viewing the sac as a static organ, the réle played by the otoliths must assume a different aspect. The fact that they are want- ing in the Brachyura, which nevertheless exhibit strong powers of orientation, might be used as an argument against their playing any important part in equilibration. But as they are present in the larval crab, and as they disappear only when the otocyst becomes highly differentiated, and when sensory hairs much more delicately constructed than the otolith bristles are developed, this argument loses most of its weight. For determining the functions of the otoliths two methods may be employed: (a) Observing the effect on equilibration and orientation following the removal of the otoliths, or the prevention of the normal process of taking them in after ecdysis. (6) Substitution of iron dust or iron filings for the otoliths, and the employment of an electro-magnet to modify the action of gravity. If the otoliths are static in function, the animals should orient themselves with reference to the resultant of the attraction of the magnet, and the pull of gravity. The first of these methods was attempted by Kreidl (’94), but failed, as he was unable completely to remove the otoliths. His results with the second method of experimentation were definite and affirmative. Lyon (’99) attempted to repeat and verify Kreidl’s work, but his experi- ments were incomplete and negative in their results. Otoliths, always normally present in macruran decapods, are lacking for only a short time after ecdysis. So short indeed is this interval, that it is extremely difficult to find otocysts of newly moulted animals which are without otoliths. Nor is it usually possible to prevent a crustacean which has been observed to cast its test, from getting new otoliths into the sac ; at least not for a sufficiently long period to allow 238 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. the animal otherwise to regain its normal condition. Even if placed at once in filtered water, some otoliths soon make their appearance, probably originating from the excreta of the animals themselves. In lobsters the larve regain their normal condition within a much shorter interval after ecdysis than do adult individuals ; their digestive tract is also much less likely to contain material suitable for the forma- tion of otoliths. Therefore, after trying in vain to completely remove the otoliths from the sacs of Crangon and Palzemonetes, my attention was directed to lobster larvee as much more favorable material than the adult shrimps. As they moult at intervals of a few days, it is also much easier to obtain them directly after ecdysis or in the very act itself. So obtained, and placed at once into filtered sea water, larve of the fourth stage may be kept without otoliths for from twenty-four to forty-eight hours, and a favorable opportunity is thus given for observing the effect produced by the lack of otoliths on the equilibrium of the animals. Observations were made on eighteen larve of the fourth and fifth stages, all of them being kept free from otoliths for at least twelve hours. Within two hours after moulting most of them swam about actively, and ate greedily when fed with bits of crab’s liver. In swimming, however, they show distinctly the phenomena manifested by shrimps which have been deprived of their otocysts. There is both “rolling” from side to side, and “pitching” forward and backward; often they swim with the ventral side uppermost. Much more easily overturned than normal larvee, they do not right themselves at once, but if turned upon the back, will continne to swim in that abnormal position. If blinded, the loss of equilibrium is still more marked. All these conditions are in strong contrast to the actions of the normal free-swimming larve of these stages, which conduct themselves in the characteristic manner already described for Palzemonetes. The observations having been made and recorded, the animals were killed, and the otocysts dissected out and examined under the micro- scope. Scarcely a particle of inorganic matter was found in the sacs of sixteen larvee. In two individuals a few small grains of sand were found in one otocyst, but the other was entirely destitute of otoliths. From the number of cases observed it seems safe to conclude that the otoliths do play an important part in equilibration, and that it is the pull of gravity upon them which stimulates the sensory hairs of the sac, If the loss of the power of accurate orientation were to be attributed to the abnormal conditions resultant upon ecdysis, it might be said in PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 239 reply that the larve were perfectly normal when observed, as far as feeding and active swimming were concerned, and furthermore that the loss of equilibration disappeared at once when a larva without otoliths was allowed to obtain them. The results of these observations are also confirmed by the following experiments. The otoliths were removed from the sacs of Palamonetes by lifting the lid which covers the aperture, and forcing a fine jet of water into the cavity. Most of the sand having been thus washed out, the animals were placed in an aquarium upon the floor of which iron filings had been scattered and were allowed to remain until the iron particles had been taken into the sac in place of grains of sand. As an electromagnet, a steel bar 8 inches long and one quarter of an inch square was used. This was ground down nearly to a point at one end; about the other end were wound many layers of fine copper wire, the termini being connected with the circuit of a small six-celled battery. The shrimps employed in the experiments (Paleemonetes) were blinded by the usual method, — painting the eyestalks with a mixture of lampblack and shellac. The pointed end of the magnet was held about 3cm. from the otocysts, at one side of and a little ventral to them. Animals with normal otoliths, if blinded, do not respond at all, and are apparently unaffected by the proximity of the magnet; they keep their normal position, dorsal side up, with the sagittal plane of the body coincident with the direction of gravity. If not blinded, they simply move slowly away from the magnet when it approaches too near. When, however, the magnet is brought into close proximity to otocysts containing iron filings, the dorsal side of the animal is turned, not toward the magnet, as might be expected if the changed position were due directly to the action of the magnet on the iron filings, but away from it. If the magnet was changed to a position on the other side of the shrimp, the turning was in the opposite direction, still away from the source of attraction. The above reaction was distinctly noted a number of times for each of the six animals experimented upon.. As Kreidl’s work was fairly com- plete, only one series of experiments was tried in confirmation of his results. When the observations had been completed, the antennules of the six shrimps were removed and the otocysts examined under the microscope. In each case particles of iron were found nearly filling the sac, and if a magnet was held close to one of the latter, the whole antennule was lifted by the attractive force, showing clearly that there must have been an effective magnetic pull upon the otoliths of the live 240 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. animals during the experiments. I believe there is only one explanation for this turning of the body away from the attracting force, and that is avery simple one. Under normal conditions the body of the shrimp is oriented with reference to gravity, and its dorso-ventral axis ap- proximately corresponds to the direction of this force. If the shrimp rotates around its chief axis either to right or left, say 90°, the direc- tion of the pull of gravity on the otoliths is at once changed, and through the medium of the latter other sensory hairs of the sac are stimulated. As a result, the shrimp turns back in a direction opposite to that in which it was rotated, until it is again in a normal relation to the direction of gravity. The employment of the magnet has no other effect than merely to change the direction of the orienting force. This is now no longer that of gravity alone, but the resultant of the two com- ponent forces, gravity and the pull of the magnet. The animal now maintains its swimming position in reference to this new line of attrac- tion, its dorso-ventral axis coincident with that line, and as a result the dorsal side is turned away from the magnet. To put it in another way, when the magnet is held close to the right side of the otocyst, the animal is stimulated precisely as it would be if rotated to the right 45°, and it responds as it would normally in righting itself, i. e., by turning its body in the opposite direction through an angle sufficient to make its dorso-ventral axis coincide with the direction of the sip a tive force ; in this case through an angle of 45°. This single series of observations completely confirm, as far as they go, the very important conclusions of Kreidl. The otoliths are found to play an important part in the functional activities of the otocyst, and _the latter is conclusively proved to be a static organ, acted upon by the force of gravity ; this force makes itself felt chiefly through the medium of the otoliths, and if they are absent, as described in a preceding set of experiments on lobster larvee, the function of the otocyst in Macrura is seriously impaired. 6. The Function of the Hairs of the Otocyst. The function of the otocyst hairs of macruran decapods which are in contact with otoliths has been already briefly discussed in.the first part of this paper. The stimulus imparted to the hair shaft through the medium of the otoliths makes itself most strongly felt at the labile base of the hair, owing to the rigidity of the shaft and the delicacy of the attaching membrane. At this point, too, the nerve fibre invariably ends, and the stimulus is thus transmitted to it, and at once carried to the PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 241 brain. In the case of adult Brachyura, however, there are no otoliths in contact with the hairs of the otocyst, consequently the effect of gravity, if not entirely null, must be at least greatly lessened, unless indeed the hairs are so differentiated as to be themselves stimulated by it. Bethe (97), acting on the idea that in tilting the animal the differ- ence in the pressure of the water might affect the hairs of the otocysts, placed crabs under very high pressures where the slight difference brought about by tilting would be practically eliminated. But he found that all the phenomena of equilibration still persisted. It is probable that in the otocyst of Carcinus the thread hairs are the most important sensory organs of the sac. The hook hairs, originally in the larva attached to otoliths, later, with the loss of the sand granules, lose much of their functional activity; the third group of hairs can- not be of great importance, as I could not demonstrate satisfactorily their nerve connections, and their structure alone is such as to preclude their being affected by very delicate stimuli. The thread hairs, how- ever, in both structure and position are fitted for the fulfillment of such a function as has been ascribed to them. The shaft is long, attenuate, only slightly fringed at the tip, and attached at the base by a very thin membrane, which allows free movement to the rigid shaft about this region as upon a joint. I have observed in studying freshly dissected otocysts that a slight tilting of the watch glass in which they were con- tained caused these hairs to sway extensively. From Clark’s experiments and my own, it was apparent that upon rotation in a horizontal plane, there was little or no compensatory movement of the eyestalks, and that when there is such a reaction, the angle of compensation is not maintained, but the eyes return at once to their original positions. Also, on rotation about the animal’s lateral axis, the angle of compensation is not as great, when the rotation is rapid and jerky, as when performed slowly and smoothly. These two facts preclude the possibility of the hairs being affected by movements of the fluid surrounding them, at least to any great extent. For if they were so affected, the angle of compensation should be the same, in what- ever plane the animals are rotated, and the position of the eyestalks should be in every case maintained by compensation movements. There still remain two ways in which the hairs may be so affected as to bring about nervous stimulus. Either they may be lighter than the surrounding fluid, and consequently tend always to float erect, no matter what position the otocyst may take relative to them; or they may be 249 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. heavier than the liquid contents of the otocyst, in which case they would be affected by gravity directly, and exposed to a greater or less pull according to their different positions in the sac. My observations made on dissections of fresh material of both young and adult crabs, do not confirm the first of these hypotheses. The hairs rarely, if ever, float upright in the fluid of the otocyst ; on the contrary they usually project out horizontally, with their tips a little lower than their bases; and such conditions would favor the second supposition, that they are heavier than the surrounding fluid. Unfortunately, when fresh material was at hand, my attention was directed toward other problems, and no dissec- tions or observations were made with the settlement of this question primarily in view. It is, however, a point well worth future experimenta- tion, for the function of these hairs is apparently similar to that of the auditory hairs of the vertebrate cristz acustice, and to clearly show how they are stimulated would throw light on an important problem in the physiology of the vertebrate ear. SUMMARY. 1. The cuticular lining of the otocyst, found in the basal segment of the antennule of all decapod Crustacea, is cast with the test at each moult. It is composed of thin chitin, and is suspended from the dorsal wall of the antennule, which presents an aperture in Macrura, in the larval stages of Brachyura, and also in adult Brachyura directly after ecdysis. 2. In Macrura a single sensory prominence is present, either on the floor or sides of the sac. In Brachyura there are three sensory regions. The sensory hairs are borne upon these cushions, usually in curved rows. 3. The otolith hairs are heavily fringed, often bent or hooked. In Macrura they are attached to the wall of the sac by a thin bulb of chitin; in Brachyura the base of the hair shaft is inserted into a cup- like depression; both methods of attachment allow the hair to sway freely upon its base. 4. The free hairs of the otocyst, found in the lobstér and all Bra- chyura, are extremely long and attenuate ; their basal attachment is deli- cate, and renders them much more sensitive than the otolith hairs. 5. All sensory hairs are formed as double-walled tubes by numerous matrix cells situated beneath the hypodermis, from which they originate. After ecdysis processes from these cells extend into the shaft of the newly formed hair. In preparation for the next moult these processes eee ee ee ee yaw OI —pmameay od nag al PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 243 are withdrawn, the matrix cells recede from the base of the old hair, and arrange themselves about the nerve fibre for the formation of the new bristle. There is a period between moults, more or less extended, dur- ing which no living substance is present in the greater part of the cavity of the hair. 6. The otoliths are grains of sand taken in froin the exterior (first, in the case of the lobster, by the fourth larva) and renewed after each moult ; they may lie free in the otocyst, or be attached to the sensory hairs. In Brachyura they are found only in the Megalops stage. 7. Glands similar in structure to the tegumentary glands are present in the lobster and crayfish beneath the sensory cushions which bear oto- lith hairs. They secrete a substance for the attachment of the otoliths to the pinnules of the bristles. 8. The innervation of the otocyst hairs and olfactory bristles is dis- tinctly unlike. (a) The otocyst hairs have each a single nerve element, and the terminal fibre ends in the enlarged base of the shaft without branching. (6) Each olfactory bristle is innervated by numerous ganglion cells (100 or more). The peripheral strand of fibres from these cells extends some distance into the cavity of the hair, terminating free and without modification of any kind. 9. The central terminations of all the otocyst fibres are in two closely connected neuropilar masses at the posterior end of the brain, median to those of the second antennz, and ventral to the optic centres. The nerve sheaths disappear as the fibres enter the “ Punktsubstanz,” and the fibrillze soon separate. They cannot be traced to determinate endings, nor are they ever directly connected centrally, with ganglion cells. 10. Each sensory nerve fibre is composed of numerous fibrille, embedded in a semi-fluid ‘ perifibrillar’ substance, which in turn is surrounded by a delicate sheath. The flowing together of the peri- fibrillar matrix causes the beaded or varicose appearances characteristic of methylen blue, and silver impregnations. The fibrillar structure can be demonstrated in both the central and peripheral portions of the fibres. 11. The sensory ganglion cells are all typically bipolar and elongate in form. They are placed at some distance from the base of the hair which they supply, and show no fibrillar structure. 12. In the shrimp-like decapods, such as Paleemonetes and Crangon, a nucleated myelin sheath surrounds each sensory fibre and ganglion 244 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. cell, extending from the neuropil of the brain nearly to the peripheral ending of the fibre. 13. Each sensory ganglion cell with its central and peripheral fibres constitutes a single nervous element or neuron. ‘The neurons are trophic units, and direct connection between two neurons was not demonstrated. 14. In those decapods which pass through free-swimming larval stages, the otocyst develops as an invagination of the dorsal ectoderm of the basal segment of the antennule, and becomes functional only at the fourth moult after hatching. 15. Invagination begins at the second larval stage, but the matrix cells which are to form the sensory hairs of the sac, make their appear- ance in the first larva, being derived from the cells of the hypodermis. 16. During the third stage the sensory hairs are formed below the floor of the shallow sac; at the next moult these become functional, the sac enlarges, and otoliths make their appearance. The otocyst is now functional, the hairs are innervated as in the adult, and more than 100 of them may be present. After the fourth stage the chief changes are the increase in the number of otocyst hairs, and the gradual constriction of the orifice of the sac. 17. In Brachyura the Zoea larva is without a functional sac. In the Megalops the otocyst is open, and contains numerous sensory hairs and otoliths. During the next two stages the aperture closes and takes on the adult condition, without otoliths. 18. Structurally, the otocyst of decapods may be compared roughly to the utriculus of such a vertebrate as Myxine; the sac of Paleemonetes to a single isolated ampulla, and its sensory cushion to a crista acustica. The closed otocyst of Brachyura has three sensory regions and is without otoliths. It therefore approaches in general structure the utriculus of the higher vertebrates. Each sensory element of the otocyst is comparable to a single sensory component in the vertebrate crista. In each there is a modified organ for the reception of stimuli, connected basally with the terminal fibre of a sensory neuron. 19. There is no part of the decapod otocyst which is structurally com- parable to the middle ear, semi-circular canals, or cochlea of vertebrates. 20. There is no direct evidence to prove that decapod Crustacea react to true sounds produced either in water or in air. The reactions formerly attributed to audition are probably due to tactile reflexes. 21. The otocyst plays little or no part in calling forth these reac- tions, and does not function as a true auditory organ. PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 245 22. Equilibration is made possible by three sets of organs, the oto- cysts, the eyes, and the tactile bristles. 23. In free-swimming decapods the otocyst is by far the most im- portant of these static organs functionally, vision being secondary to it. Four facts go to prove this: (a) The removal of the otocysts causes a much greater loss of power of orientation, and a greater decrease in the compensatory movements of the eyestalks, than the loss of vision. (5) Decapods and Entomostraca normally without otocysts either swim in unstable equilibrium, or in a position identical to that which an inanimate object of the same form and weight would take under the influence of gravity. (c) Lobster larve without functional otocysts are unstable in their swimming movements, but orient themselves with great accuracy at the stage when the sac becomes an active sense organ. (d) If iron filings are substituted for the otoliths, and an electro- magnet is employed to modify the effect of the pull and direction of gravity, shrimps orient themselves with reference to the direction of the resultant pull of the two forces, precisely as they do to the attraction of gravity alone. 24. In lobster larvee of the third and fourth stages there is a direct correlation between the metamorphosis of the otocyst from a func- tionless to an active organ, and the changes in the swimming position of the animal. When the sac is inactive (third stage), the swimming position of the body and appendages is an adaptation which places the larve in comparatively stable equilibrium. As the otocyst becomes functional (fourth stage), this adaptation is no longer necessary, and a much less stable position is maintained, but one more favorable for rapid locomotion. 25. The otoliths in Macrura and larval Brachyura are the means by which the pull of gravity is transmitted to the hairs of the otocyst. On their complete removal there is loss of equilibration and power of orientation ; if iron filings are substituted for them, shrimps may be made to respond to the attraction of an electromagnet. 26. In adult Brachyura otoliths are normally lacking. The otolith hairs have become practically functionless, and the thread hairs are modified in such a way as to make them directly responsive to the attraction of gravity without the aid of the otoliths. VOL. XXXVI. — NO. 7 6 246 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. BIBLIOGRAPHY. fElianus, C. 1784. De natura animalium. Libri 17, Lipsie. Apathy, S. °97. Das leitende Element des Nervensystems und seine topographischen Beziehungen zu den Zellen. Mitth. Zool. Stat. Neapel, Bd. 12, Heft 4, pp. 495-748, Taf. 23-32. Bate, C. S. °55. On the Homologies of the Carapace, and on the Structure and Function of the Antenne in Crustacea. Ann. Mag. Nat. Hist., Ser. 2, Vol. 16, pp. 36-46, Pls. 1-2. Bate, C. S. 58. On the Development of Decapod Crustacea. Phil. Trans. London, Vol. 148, pp. 589-605, Pls. 40-46. Beer, T. 98. Vergleichend-physiologische Studien zur Statocystenfunction. I. Ueber den angeblichen Gehérsinn und das angebliche Gehérorgan der Crustaceen. Arch. f. ges. Physiol., Bd. 78, pp. 1-41. Beer, T. 99. Vergleichend-physiologische Studien zur Statocystenfunction. II. Ver- suche an Crustaceen (Peneus membranaceus.) Arch. f. ges. Physiol., Bd. 74, pp. 364-382, 2 Figg. Bethe, A. 794. Ueber die Erhaltung des Gleichgewichts. Biol. Centralbl., Bd. 14, pp. 95-114, 563-582. Bethe, A. 94a, Vergleichende Untersuchungen iiber die, Functionen des Central- nervensystems der Arthropoden. Arch. f. ges. Physiol., Bd. 68, pp. 449- 545, Taf. 3. Bethe, A. 95. Studien iiber das Centralnervensystem von Carcinus menas nebst An- gaben iiber ein neues Verfahren der Methylenblaufixation.” Arch. f. mikr, Anat., Bd. 44, pp. 579-622, Taf. 34-36. Bethe, A. °95a, Die Otocyste von Mysis. Zool. Jahrb., Abth. f. Anat., Bd. 8, pp. 544— 564, Taf. 37. q PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 247 Bethe, A. 797. Das Nervensystem von Carcinus menas. Arch. f. mikr. Anat., Bd. 50, pp. 460-516, 589-639; Taf. 25-30, 33. Bethe, A. 798. Das Nervensystem von Carcinus menas. Arch. f. mikr. Anat., Bd. 51, pp. 382-450, Taf. 16, 17. Braun, M. 75. Ueber dic histologische Vorgange bei der Hautung von Astacus fluvia- talis. Arbeit. Zool.-Zoot. Inst. Wurzburg, Bd. 2, pp. 121-166, Taf. 8-9. Bunting, M. 793. Ueber die Bedeutung der Otolithenorgane fiir die geotropischen Functio- nen von Astacus fluviatalis. Arch. f. ges. Physiol., Bd. 54, pp. 531-837. lark, J. P. 96. On the Relation of the Otocysts to Equilibrium Phenomena in Gela- simus pugilator and Platyonichus ocellatus. Jour. of Physiol., Vol. 19, pp- 327-343, 5 text figs. Claus, C. 75. Entwicklung, Organisation, und systematische Stellung der Arguliden. Zeitschr. f. wiss. Zool., Bd. 25, pp. 247-284, Taf. 14-18. Claus, C. 791. Ueber das Verhalten der nervésen Endapparates an den Sinneshaaren der Crustaceen. Zool. Anzeiger, Jahrg. 14, pp. 363-368. melage, -Y. 87. Sur une function nouvelle des otocystes comme organs d’orientation locomotrice. Arch. Zool. exp. et gen., Sér. 2, Tome 5, pp. 1-26. Engelman, T. W. 87. Ueber die Function der Otolithen. Zool. Anzeiger, Jahrg. 10, pp. 439- 444. Farre, A. 43. On the Organ of Hearing in Crustacea. Phil. Trans. London Vol. 1843, pp. 233-242, Pls. 9-10. Frey, H., und Leuckart, R. *47. LBeitrage zur Kenntniss wirbelloser Thiere, mit besonderer Beriicksichiti- gung der Fauna des Norddeutschen Meeres. 170 p., 2 Taf., Braunschweig. Garbini, A. 80. Sistema nervosa del Palemonetes varians. Atti Soc. Veneto-Trentina, Padova, Tome 7, pp. 179-199, Tab. 13-18. Goode, G. B. *78. The Voices of Crustaceans. Proc. U. S. Nat. Mus., Vol. I, pp. 7-8. Haeckel, E. *57. Ueber die Gewebe des Flusskrebses. Arch. f. Anat. u. Physiol., Jalirg. 1857, pp. 469-568, Taf. 18-19. 248 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Hensen, V. °63. Studien iiber das Gehérorgan der Decapoden. Zeitschr. f. wiss. Zool., Bd. 18, pp. 319-402, Taf. 19-22. Hensen, V. 99. Wie steht es mit der Statocysten-Hypothese ? Arch. f. ges. Physiol., Bd. 74, pp. 22-42. Herrick, F. H. °95. The American Lobster: A Study of its Habits and Development. Bull. U.S. Fish. Comm. for 1895, xi + 252 p., Pls. A-J, 1-54. Huxley, T. H. 51. On the Auditory Organs in the Crustacea. Ann. Mag. Nat. Hist., Ser. 2, Vol. 7, pp. 304-306, 373-374, Pl. 14. Huxley, T. H. °80. The Crayfish: An Introduction to the Study of Zoology. xiv +371 p., 82 text figs. New York. Jourdain, S. °80. Sur les cylindres sensoriels de ’antenne interne des Crustaces. Compt. Rend. Acad. Sci., Paris, T. 91, pp. 1091-1098. Kreidl, A.’ °93. Weitere Beitrage zur Physiologie des Obrlabyrinthes. (II. Mitth.) Versuche an Krebsen. Sitzungsber. Akad. d. Wiss. Wien, Bd. 102, math.-naturw. Cl., Abth. 3, pp. 149-174, Taf. 1-2, 5 Holzschnitten. Kroyer, H. °59. Forség til en monographisk Fremstilling af Kraebsdyrslaegten Ser- gestes. Med Bemerkninger om Decapodernes Horeredskaber. Dansk. Vidensk. Selskab. Skrift., R. 5, Bd. 2, pp. 219-302, Taf. 1-5 a. Langdon, F. E. °95. The Sense-organs of Lumbricus agricola, Hoffm. Jour. Morph., Vol. 11, pp. 193-232, Pls. 13, 14. Langdon, F. E. :00. The Sense-organs of Nereis virens, Sars. Jour. Comp. Neur., Vol. 10, pp. 1-77, Pls. 1-3. Lemoine, V. °68. Recherches pour servir a l’histoire des systémes nerveux musculaire et glandulaire de l’Ecrevisse. Ann. Sci. Nat., Sér. 5, Zool., T. 9, pp. 99- 280, Pls. 6-11. Leuckart, R. 53. Ueber die Gehérwerkzeuge der Krebse. Arch. f. Naturg., Jahrg. 19, Bd. 1, pp. 253-265. Leuckart, R. 59. Ueber die GehGrorgane der Decapoden. Arch. f. Naturg., Jahrg. 25, Bd. 1, pp. 265-266, Taf. 7. Leuckart, R., und Frey, H. See Frey, H., and Levckanrt, R. = 1 4 tee ey ah ek 2s Stee agin de PRENTISS: THE OTOCYST OF DECAPOD CRUSTACEA. 249 Lewis, M. 98. Studies on the Central and Peripheral Nervous Systems of two Poly- chete Annelids. Proc. Amer. Acad. Arts and Sci., Vol. 33, pp. 223- 268, 8 pls. Apr. 1898. Leydig, F. 55. Zum feineren Bau der Arthropoden. Arch. f. Anat. u. wiss. Med., Jahrg. 1855, pp. 376-480, Taf. 15-18. Leydig, F. °57. Lehrbuch der Histologie der Menschen und der Thiere. xii+551 p., Frankfurt a. M. Leydig, F. °60. Ueber Geruchs- und Gehérorgane der Krebsen und Insecten. Arch. f. Anat. u. wiss. Med., Jahrg. 1860, pp. 265-314, Taf. 7-9. Lyon, E. P. 99. A Contribution to the Comparative Physiology of Compensatory Move- ments. Am. Jour. Physiol., Vol. 3, pp. 86-114, Nov. 1899. Milne-Edwards, H. °57-81. Lecons sur la physiologie et l’anatomie comparée de l’homme et des animaux. ‘T. 1-14, Paris. Minasi, A. 1775. Dissertazione seconda su de’ timpanetti dell’ udito scoverti nel gran- chio paguro, e sulla bizzarra di lui vita. 136 p. Napoli. Morrill, A. D. 798. The Innervation of the Auditory Epithelium of Mustelus canis, De Kay. Jour. Morph., Vol. 14, pp. 61-82, Pls. 7-8. Miller, J. 737. Handbuch der Physiologie des Menschen. Bd. 2, Coblenz. Parker, G. H. 790. The Histology and Development of the Eye in the Lobster. Bull. Mus. Comp. Zodl., Vol. 20, No. 1, pp. 1-60, Pls. 1-4. Parker, T. 78. Note on the Stridulating Organ of Palinurus vulgaris. Proc. Zool. Soe. London, 1878, pp. 292; 442-444, Pl. 30. Rath, O. vom. 87. Ueber die Hautsinnesorgane der Insecten. Zool. Anzeiger, Jahrg. 10, pp. 627-631, 645-649. Rath, O. vom. °88. Ueber die Hautsinnesorgane der Insekten. Zeitschr. f. wiss. Zool., Bd. 46, pp. 413-454, Taf. 30-31. Rath, O. vom. ‘ 91. Zur Kenntniss der Hautsinnesorgane der Crustaceen. Zool. Anzeiger, Jahrg. 14, pp. 195-200, 205-214. 250 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Rath, O. vom. 94. Ueber die Nervenendigung der Hautsinnesorgane der Arthropoden nach Behandlung mit der Methylenblau- und Chromsilbermethode. Berichte Naturf. Gesellsch. Freiburg, Bd. 9, Heft 2, pp. 1-28, Taf. 2. Reichenbach, H. °86. Studien zur Entwicklungsgeschichte des Flusskrebses. Abhandl. Senckenberg. naturf. Gesellsch. Frankfurt a. M., Bd. 14, 137 p., 14 Taf. Retzius, G. °90. Zur Kenntniss aes Nervensystems der Crustaceen. Biol. Untersuch., Bd. 1, pp. 1-50, Taf. 1-14, Stockholm. Retzius, G. *92. Ueber die neuen Prinzipien in der Lehre von der Einrichtung des sensiblen Nervensystems. Biol. Untersuch., Bd. 4, pp. 49-56, 9 Figg. Stockholn. Retzius, G. °94. Zur Entwicklung der Zellen des Ganglion spirale acustici, und zur En- digungsweise des Gehdrnerven bei den Satigethieren. Biol. Untersuch., Bd. 6, pp. 52-57, Taf. 24-25. Retzius, G. 95. Das sensible Nervensystem der Crustaceen. Biol. Untersuch., Bd. 7, pp. 12-18, Taf. 4-6. Rosenthal, C. "11. Ueber die Geruchssinne der Insecten. Reil’s Arch. f. Physiol., Bd. 10, pp- 427-439, Taf. 8. Sars, G. O. °67. Histoire naturelle des crustaces de ]’eau douce de Norwége. iii+145 p., 10 pls. Christiana. Siebold, C. T. von. 44. Ueber das Stimm- und Gehororgan der Orthopteren. Arch. f. Naturg., Jahrg. 10, Bd. 1, pp. 52-81, Taf. 1. Siebold, C. T. von. 48. Lehrbuch der vergleichenden Anatomie der wirbellosen Thiere. xiv + 679 pp., Berlin. Souleyet. 43. Observations anatomiques, physiologiques, et zoologiques sur les mol- lusques ptéropodes. Compt. Rend. Acad. Sci. Paris, T. 17, pp. 662-675. Treviranus, G. R. 02-22. Biologie, oder Philosophie der lebenden Natur, fiir Naturforscher und Aerzte. 6 Bde., Gottingen. Verworn, M. z ; 91. Gleichgewicht und Otolithenorgan. Arch. f. ges. Physiol., Bd. 50, pp. 423-472. PRENTISS : All Figures were outlined with an Abbe camera lucida. EXPLANATION THE OTOCYST OF DECAPOD CRUSTACEA. 251 OF THE PLATES. Tube length was usually 160 mm., with projection distance to the table, 410 mm. The magnifi- cations are given with the descriptions of the several figures. Drawings were made from sections, unless the contrary is stated. The orientation of the figures is given for each plate. cl. gl. Clg.) cl. ma. . cont. erc’oes. crs. Sns. sts. UT «= for. ass. for. ¢. Yoram: Sor. pi’ph. . ‘ae glib. . gn. olf. . Wdrm. . lab.@. . . iab. p. ae mal.. .« mb. sph. n. at. n.at.g . wl. . n’lem ABBREVIATIONS. Gland cell. nl. tu. Ganglion cell. n. opt. Matrix cell. n, ot. Circumeesophageal con- n’pil. at. . nective. n’pil.at.2 . Sensory crista. n’pil. opt. . Cuticula. n. teq. New cuticula. of. Duct. ot’cy. Fibrillations. ot Ith. Association nerve fibre. Central nerve fibre. Nerve fibre. Peripheral nerve fibre. Flagellum. Globulus. Olfactory ganglion. Hypodermis. Anterior lip of orifice to otocyst. Posterior lip of same. Lumen of otocyst. Hammer. Spherical membrane. Antennular nerve. Nerve of 2d antenna. Neuroblasts. Neurilemma. pinn. pr’c. pr’ pl. rm. l. ro. SOU, sts set. fil. . set. l. set.m. .- set. olf. . set. ot. set. p. set.ta. . set. tac. tb. set. . TCE os tu. myl. Sheath nucleus. Optic nerve. Otic nerve. Neuropil of antennule. Neuropil of antenna. Optic neuropil. Tegumentary nerve. Orifice. Otocyst. Otoliths. Pinnules of hairs. Protoplasmic process. Lateral branch of an- tennular nerve. Rostrum. Group hairs. Thread hairs. Lateral hairs. Median hairs. Olfactory hairs. Otic hairs. Posterior hairs. Hook hairs. Tactile hairs. Hair tube. Tectum of otocyst. Myelin sheath. PRENTIss. — Otocyst Crustacea, PLATE 1. All Figures are of Palemonetes. In Figure 1 anterior is up on the plate ; in Figures 2, 8, and 4 the dorsal side is up, and the anterior end in Figure 4 is at the left. Lines numbered 2, 3, 4,5 in Figure 1 indicate the planes of the sections shown in Figures 2, 3, 4, and 5 respectively. Fig. 1. Dorsal view of the basal segment of both antennules, showing otocysts and the arrangement of the hairs in the sac. X 30. Fig. 2. Somewhat oblique transverse section, extending from dorsal anterior to ventral posterior, of both antennules and the rostrum, through the posterior ends of the otocysts (compare line 2, Fig. 1). X 64. Fig. 3. Transverse section of right antennule through the orifice of the sac, show- ing tectum and otoliths (compare line 3, Fig. 1). X 64. Fig. 4. Parasagittal section through right antennule and brain, showing the course of the otic nerve, with a single nerve element drawn diagram- matically (see line 4, Fig. 1). X 64. ips IP o 3 p qeeteers e PRENTiss. — Otocyst Crustacea. PLATE 2. All Figures are of Palaemonetes. The dorsal side is up in Figure 5, and the an- terior end at the right. Fig. 5. Fig. 6. Fig. 7. Fig. 7a. Fig. 8. Fig. 9. Parasagittal section through the lateral side of the right otocyst (see line 5, Fig. 1). x 64. Transverse section through the posterior end of the sensory cushion, showing two lateral hairs, the base of a median one and a group of ganglion cells. XX 168 Otocyst hair and matrix cells. X 600. Another otocyst hair and matrix cells. x 600. Sensory hair of the otocyst and the ending of its peripheral nerve fibre. x 600. Fundament of developing sensory hair from an abdominal exopod, showing matrix cells about the nerve fibre. Methylen-blue prepara- tion. X 600. NTIsS—OTOCYST GRUSTACEA. B. Meisel, lith. Beston =) Pvt PRENTISS — Otocyst Crustacea. PLATE 3. All Figures are from methylen-blue preparations of Palwmonetes. Anterior is up on plate in Figure 12. 7 Fig. 10. Part of abdominal exopod showing tubes of developing tactile hairs and their innervation. X 126. Fig. 11. Peripheral nerve endings in the tactile hairs of the second maxilliped. x95; ‘ Fig. 12. Dorsal view of antennules and brain, showing sensory neurons and cen- tral endings of the otocyst nerve fibres. The peripheral endings in the left otocyst are diagrammatic. X 30. Paeviss— CEs ee CRUSTACEA La ta | ecu j/fll! =s\-- ¥- COnLTC.OES . = xX Ng PRENTISS. — Otocyst Crustacea. ‘ PLATE 4. Figures 13-18 are of Palemonetes. Figures 19-22 are of lobster larve. In Figures 19-21 dorsal side is up and the lateral side is at the right; in Figure 22 anterior is at the right, dorsal side up. Fig. 13. Portion of inner flagellum of first antenna, showing olfactory hairs and Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 14. 22. their peripheral ganglionic masses. Methylen blue. X 125. Gustatory hairs and nerve elements from the basipod of second maxilla. Methylen blue. x 95. Fibrillations in peripheral otic nerve fibre. Methylen blue. X 1800. Ganglion cell of otocyst and peripheral nerve process surrounded by- myelin sheath. x 600. Ganglion cell of otocyst, and sheath nucleus. X 1300. Ganglion cell of otocyst. Methylen blue. X 770. Transverse section through right antennule of first lobster larva. X 168. Transverse section through right antennule of second lobster larva; beginning of invagination. X 168. Transverse section through right antennule of third lobster larva. xX 168. Parasagittal section through antennule of second lobster larva. XX 168. ee ote aed Faarest ees * en eee —— ‘ese ote Se er bn of "*t 00 © Q . PRENTIsS. — Otocyst Crustacea. PLATE 5. All Figures are of lobster Jarve. In Figures 23 and 25 dorsal is up, and anterior at the right ; in Figure 24 dorsal is up and lateral is at the right; anterior is at the right in Figure 26. Fig. 23. Fig. 24. Fig. 25. Fig. 26. Parasagittal section through the antennule of the third larva. X 168. Transverse section through the posterior part of the right otocyst in fourth larva. X 168. Parasagittal section of the same. X 128. Diagrammatic dorsal view of floor of right sac dissected out to show arrangement of the hairs and their innervation. Methylen blue. xX 168. Developing hairs of the otocyst in the third larval stage. X 600. AT = Pee eet me Fe dese? 95 lS al : we E 4 = PRENTIss. — Otocyst Crustacea. PLATE 6. Crangon ; in Figure 28 dorsal is up; in Figure 29 anterior is up. Fig. 28. Transverse section of both antennules through the sensory cushions of the otocysts. X 965. Fig. 29. Reconstruction from ten frontal sections through the base of both anten- nules and the brain. A semi-diagrammatic -nerve element is shown at the left in black. X 965. 28 29 Oe ren s peaunad . : i geet rhe Fee Ae A ee ae gat hriglesay . —— Pees 4 pes vyaee ee ye _ _ Does ~~ PRENTIss. — Otocyst Crustacea. PLATE 7. Figures 30 and 31 are of Crangon. All others are of Cambarus. In Figure 30 dorsal is up and lateral is at the left ; in Figure 36 dorsal is up and lateral is at the right. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 30. 31. 32. 33. Parasagittal section of antennule and brain. X 95. Otocyst hair. X 600. Olfactory hairs from basipod of second maxilla, showing innervation. Methylen blue. X 95. Posterior row of otocyst hairs and their nerve elements. Methylen blue. X 95. Tactile hair from scaphognathite of second maxilla, showing innervation. Methylen blue. X 985. Transverse section through shaft of otocyst hair. X 770. Diagrammatic transverse section of the right antennule through the pos- terior part of the otocyst. X 15. ties ana Ld \J ! j V\ A | set.tac.¥ | LIA AY LU 1 1A A ie 7 iA, \ /\} ‘ 2. J Y "s f ; J | Gs ue an Vy (y,. torn: ———- | } Clg... a ona eiabeae | | BS. | PRENTIss. — Otocyst Crustacea. PLATE 8. All Figures are of Cambarus. Dorsal side is up in Figures 37 and 38; in Figure 38 anterior is at the left; anterior is up in Figure 40. Fig. 37. Transverse section through the orifice of both otocysts ; left antennule is diagrammatic. X 25. Fig. 38. Parasagittal section through right antennule and brain. X 25. Fig. 39. Tegumentary gland from the sensory cushion of the otocyst. X 600. Fig. 40. Dorsal view of the sensory cushion of the left otocyst dissected out, showing the arrangement and innervation of the hairs. Methylen blue. X 62. PRENTIss. — Otocyst Crustacea. PLATE 9. Figure 41 is of Cambarus; all others are of Carcinus. In Figures 41 and 46 anterior is up; in Figures 42-45 dorsal is up and anterior at the left; in Figures 47 and 48 dorsal is up and lateral is at the right. Fig. 41. Ventral view of brain, showing central endings of otic and antennular nerves. Methylen blue. X 30. Figs. 42-45. Outlines of four parasagittal sections through the left otocyst of Carcinas, cut along the lines of section marked with corresponding numbers in Figure 46. Figure 45 is most median, Figure 42 most lateral in position. X 15. Fig. 46. Dorsal view of both antennules. Numbered lines (42-48) indicate planes of section of corresponding Figures. X 8. Fig. 47. Transverse section through the orifice of right otocyst (see Fig. 46, line Liye Seb Fig. 48. Transverse section through anterior end of the otocyst (see Fig. 46, line 48). X 165. = Cr , — ya —OTocYST DECAPOD CRUSTACEA ( RUSTACEA. PLATE 9 lab.p. 42. ad’, 45. Fig. Fig. Fig. Fig. Fig. Fig. Fig. PRENTISS. — Otocyst Crustacea. 49. 50. 51. 52. 53. 54. 56. PLATE 10. All Figures are of Carcinus ; anterior is up in Figure 56. Group hair. XX 600. Transverse section through the sensory cushion of the hook hairs. x 168. Hook hair. X 600. Portion of outer flagellum of antennule, showing the bases of olfactory hairs and their innervation. X 95. Thread hairs and their nerve elements. Methylen blue. X 96. Tip of thread hair. X 1300. Nearly frontal section, inclining dorsal and forward, through both anten- nules and brain. X 25. PLATE 10. B. Meisel, lith. Boston. - Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE. Vou. XXXVI. No. 8. ON A COLLECTION OF BIRDS FROM THE LIU KIU ISLANDS. By Outram Banes. CAMBRIDGE, MASS., U.S. A.: PRINTED FOR THE MUSEUM. Juty, 1901. 7 . No. 8.— On a Collection of Birds from the Liu Kiu Islands. By Outram Bancs. THe Museum has recently acquired from Mr. Alan Owston of Yoko- hama an interesting collection of birds from the Yayeyama, or southern group of the Liu Kiu Islands. Though consisting of but one hundred and seven specimens, comprising fifty-six species, it contains six forms apparently hitherto undescribed. The collection was made by Mr. Ishida Zensaku and assistants from February to July, 1899, mostly in the Island of Ishigaki; some of the species were taken in the islands of Taketomi, Kobama, Hamarlijima, Kuroshima, Hatojima, and Iruduroto. The systematic sequence adopted is that of Stejneger in his Catalogue of Birds hitherto recorded from the Liu Kiu Islands.1_ I am indebted to the Museum authorities for placing the collection at my disposal for study, and am under special obligation to Dr. Leonhard Stejneger of the United States National Museum. Dr. Stejneger has made extensive studies of the fauna of the Liu Kiu Islands, and his aid and advice in comparing the specimens of the present collection with those in the National Museum have been of great value. Jam also indebted to Mr. E. W. Nelson of the Biological Survey for comparing the noddy and sooty terns with those in the Department of Agriculture collection. In the following descriptions all measurements are in millimetres; the wing is measured in its natural curve, and not flattened down on the rule; the tail is measured by thrusting one point of the dividers to the base of the tail feathers and measuring thence to the tip of the longest rectrix. All colors, when definitely expressed, are according to Ridgway.” Sterna melanauchen Temm. Two specimens, adult ¢ and adult 9, from a small island near Taketomi, were taken June 20. [Eggs were collected from June 25 to July 5; a single egg laid on the ground. | 1 Proc. U. S. Nat. Mus., 1887, Vol. X. pp. 414-415. 2 Ridgway, R. A Nomenclature of Colors for Naturalists, etc. Boston, 1886. 3 A list of the Zensaku collection, containing many notes on the distribution, nesting habits, etc., of the species taken, was published by Mr. Alan Owston (Yo- kohama, 1899). In this paper extracts from Owston’s list are in brackets. VOL. XXXVI. — NO. 8 256 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Sterna dougalli gracilis (Gourp). Two specimens, an adult ¢ and an adult 9, taken June 7 on a small island near Ishigaki. [Eggs were collected from June 19 to July 5.] These speci- mens are extreme of the slender-billed small form to which Gould’s name gracilis applies. Specimens from western Europe and Africa agree closely in measurements with those from eastern North America and the West Indies. The red bill claimed as a character of gracilis may be due to age, many young specimens from America having red bills, while in the adult birds it is black. The differences between the two races of the Roseate Tern in size and in measure- ments of the bill are well marked. The Liu Kiu Islands specimens agree in measurements with the Australian form, upon which Gould based his S. gracilis, and there can be no doubt of their identity. The measurements of the two specimens are as follows : — No. Sex. Wing. Tail.1 Tarsus. Exposed Culmen. 387,504 oa 221 110 20.2 36.6 387,305 Q 216 109 19.4 36. Sterna fuliginosa crissalis (Bamp). Two specimens from a small island near Iriomote, adult # and adult 9, taken June 10. [Eggs, one in a clutch, laid on the rock, were taken June 1.] Sterna bergii boreotis,? subsp. nov. Type. — Mus. Comp. Zo6l., No. 37,301. A single adult @ in full breeding plumage from Ishigaki, June 15, 1899. [Said to breed on Ishigaki.] Subspecific Characters. — As small as the pale gray Sterna bergii poliocerea of Tasmania and South Australia ; differing from it in having the wings, tail, and mantle very dark smoke gray, almost mouse gray. Color. — Type, adult @ in full plumage. Forehead, cheeks, lores, ear-coverts, neck all round, and whole under parts, including lining of wing and bend of wing, pure white ; crown and long occipital crest glossy black ; mantle, wings, rump, upper tail coverts, and upper surface of middle rectrices dark smoke gray, darkest on wings and middle of back, where the color is almost mouse gray ; primary quills white; Ist primary with outer web, a band along quill on inner web and tip blackish, with a silvery suffusion which is most marked toward centre of feather; broad outer margin of inner web, below the black tip 1 The tails are measured to the end of the second rectrix, the streamer varying too much in length individually to be taken into account. ‘ 2 Boreotis, northern. eieeela beh an Bl wren nt: Pa eed ie OF oe BANGS: BIRDS FROM THE LIU KIU ISLANDS. 257 white ; 2nd primary similar but black tip deeper in color and extending a short distance down outer margin of inner web, thus enclosing the white of inner web for a short distance; 3rd, 4th, and 5th primaries like 2nd, but black tip gradually growing deeper in color; outer rectrices above pale smoke gray at tips and along shafts, pale grayish white toward base; 2nd and 3rd rectrices darker on the outer webs and at tip and whitish toward base of inner webs ; bill, in dried specimen, dull yellow clouded with olive toward base; feet and tarsi blackish. Measurements1— Adult ¢@, type, wing 344; tail 178; tarsus 28; culmen 62. Remarks. — Sterna bergit was first recorded from this region (breeding on small islands off the north cuast of Formosa) by Swinhoe (Ibis, Vol. II. p. 68, 1860); since then two specimens have been noted by Stejneger, both from the Yayeyama Islands, the first in Proc. U. S. Nat. Mus., 1887, Vol. X. p. 392; the second in Vol. XIV. p. 490, 1891. But the question Stejneger raised in 1887, “Will anybody kindly inform me what name properly belongs to the smaller dark birds from the China seas? ” has hitherto remained unanswered. My type of Sterna bergit boreotis agrees with the descriptions of Stejneger’s specimens, and I propose for the small dark northern form of Bergius’s tern the trinomial given above. When Saunders wrote his account of Bergius’s tern, he hada large series of specimens at his command. He devotes but a few lines to the exceedingly interesting geographical variations of this wide-spread species, and after pointing out, in rather a vague way, how well marked the various races are, ends by including them a!l under one name. The principal races of Sterna bergii may be indicated as follows : — 1. Sterna bergii bergii Licht., South Africa, large, gray of upper parts pale. 2. S. bergit velox (Cretzschm), Red and Arabian Seas and Bay of Bengal, large, gray of upper parts very dark. 3. S. bergii pelecanoides (King), northern parts of Australia, intermediate between the last two in size and coloration. 4. S. bergii poliocerca} (Gould), Tasmania and South Australia, small, gray of . upper parts pale. 5. SS. bergii boreotis, subsp. nov., Liu Kiu Islands and Northern China Sea, small, gray of upper parts very dark. Still another race that may prove distinct is the Polynesian S. rectirostris Peale, described from the Fiji Islands. 1 Three specimens of S. bergii poliocerca in the Mus. Comp. Zool. afford the following measurements : — No. Sex. Locality. Wing. Tail. Tarsus. Culmen. 8,781 cs Australia. 334 158 31 59.5 12,018 g (2) Melbourne, Aust. 332 173 27 56. 8,782 d (2) Australia. 340 146 30 59. For further measurements, see Stejneger, Proc. U. 8. Nat. Mus., 1887, Vol. X. pp. 393-394. 8 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. bo Ol Anous pullus,! sp. nov. Type. — Mus. Comp. Zodl., No. 37,298. Two specimens, an adult ¢ and an adult 9, from a small rocky island near Iriomote, June 10. [Eggs, one in a clutch, laid on the bare rock, were taken July 1.] Characters. — A large very dark brown noddy with a gray crown, nearest to A. rousseaut Hartl. of Madagascar and adjacent islands, from which it differs by being much darker in color and slightly smaller in size. Color. — Adults, in unworn, full breeding plumage. Narrow superciliary streak, ending above eye, lower eyelid, and a spot on upper eyelid whitish ; forehead pearl gray, this color extending over crown and gradually darkening to slate gray on occiput, and thence merging on hind neck into the brown of upper parts; lores and region above the eye below the whitish streak black; upper parts rich dark chocolate brown, with a slight grayish cast; primaries and rectrices dark blackish brown; chin and sides of head blackish slate ; rest of under parts deep chocolate brown ; lining of wing brownish slate ; bill, in dried specimens, black ; feet and toes reddish brown. Measurements : — No. Sex. Wing. Tail. Tarsus. Culmen. 37,297 g Topotype. 273 164.5 25. 39. 37,298 9 Type. 271 159.5 24.5 38. Remarks. — A comparison of the two specimens upon which I base this new noddy with the material in the National Museum and the Museum of Com- parative Zodlogy shows them to be much nearer to A. rousseaut than to any of the other forms. The comparison was made with skins of A. rousseaut from the Seychelles and Mauritius. The Liu Kiu birds are much darker in color throughout, especially so about chin, sides of neck, and breast, and they are also smaller, the wing of the Mauritius specimen being 285 mm. long, and I have no hesitation in proposing a name for the Liu Kiu noddy. Compared with other noddies, the differences are still greater ; thus the Liu Kiu form is much darker than A. ridgwayi Anthony from Socorro and Tres Marias, especially about sides of head and throat, and the crown is darker and grayer. From A. galapagensis Sharpe the new species differs in not having so black a body or such a dark gray crown. From the noddy of eastern America—true A. stolidus —the Liu Kiu bird is very distinct, and can at once be told by its larger size and gray crown and forehead, the forehead and most of the crown in A. stolidus being white or yellowish white. A. pullus differs much from the small slender-billed species, A. lewcocapillus, A. hawaiiensis, and A. tenuirostris, in being larger and having a stouter bill. 1 Pullus, dark-colored, dusky. BANGS: BIRDS FROM THE LIU KIU ISLANDS. 259 Puffinus leucomelas Temm. Two specimens from a small island near Iriomote, taken June 7. [One egg was taken July 1, from a hole in the rock about six feet deep. ] Bulweria bulweri (Jarp. & SELB.). One adult 9 from Hanarejima, June 25. [Two eggs supposed to belong to this species were taken on the.same island, June 20. ] Arenaria interpres (Liny.). One adult @ in full plumage, Ishigaki, May 10. Charadrius dominicus fulvus (Gmet.). Two females from Ishigaki, March 1 and June 1. Aegialitis alexandrina (Liyw.). One specimen, March 13, Ishigaki. [Eggs were collected, April 29 to June 20.] Ochthodromus mongolus (PAtt.). One 9 from Ishigaki, June 1. Actitis hypoleucos (Lryx.). One 9 from Ishigaki, March 12. Heteractitis brevipes (VI=ILt.). One 9 in winter plumage, Ishigaki, March 12. Gallinago gallinago (Liny.). One 9 from Ishigaki, March 25. Limnobzenus pheopygus (Srersy.). Three specimens from Ishigaki, adult ¢ taken May 1, adult 9 June 20, and achick June 19. The chick is covered with black down, which on the back is shining blue black, the bill and a patch of bare skin below the eye are yellow. 260 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. The wing in the adult ¢ is 105.4, in the adult 9 104. Neither of these has white spots on the outer web of Ist primary, such as Stejneger describes. [Nests containing six eggs each were found among reeds from April 10 to — July 4.] Rallina sepiaria (SrEuy.). Two adults from Ishigaki, # taken March 20 (wing 146), 9 taken April — 2 (wing 150). ® Gallinula chloropus orientalis /Horsr.). Two adults, ¢ and 9, from Ishigaki, taken March 21. Fuligula fuligula (Liyy.). Two adults from Ishigaki, 9 taken May 20, g June 13. The male lacks the white spots on the chin. Anas zonorhyncha Swiyz. Two adults from Ishigaki, g May 10, Q June1. [Many nests were found placed on the ground among grass, and eggs, seven in a clutch, taken from April 19 to June 25.] Nettion crecca (Liyy.). One female from Ishigaki, March 7. Dendrocygna! javanica (Horsr.). Two adults from Ishigaki, ¢ taken May 25, 9 June l. [Nests were found on the ground among tall grass, and eggs, six in a clutch, taken from May 31 to June 21.] Sula sula (lryvy.). Two specimens, adult # from Iriomote, June 20, adult ? from Ishigaki, June 15. [Eggs were found two in a clutch, on outlying rocks, May 12 to June 13.] Gorsachius melanolophus (Rarrtes). Two adults from Ishigaki, # March 23, Q June 7. 1 This name is by many ornithologists improperly spelled, “ Dendrocycna.” Swainson’s original spelling was “ Dendrocygna.” BANGS: BIRDS FROM THE LIU KIU ISLANDS. 261 Demiegretta ringeri Sresn. One fine adult female, taken in Ishigaki, March 25. This skin agrees with Stejneger’s description, and the northern reef heron is a valid form, differing, as pointed out by Stejneger, from the southern reef heron in its gray head and occipital crest. It is, however, not recognized by Sharpe in the Catalogue of Birds in the British Museum. Nannocnus eurythmus (Swiyz.). Two adults from Ishigaki, ¢ taken March 25, 9 June 10. [Nests built in reeds about two feet from the ground, containing six eggs each, were found from May 19 to July 3.] Pyrrherodias manillensis (MEYEy). Six specimens, all from Ishigaki, adult ¢ June 20, adult 9 May 20, and four nestlings June 1. [Eggs were taken from April 22 to May19. The nests were placed on oak and other trees, at from 20 to 30 feet from the ground, and usually contained four eggs each. ] This heron was first recorded from the Yayeyama Islands by Stejneger in 1891, who doubtfully referred! it to Ardea purpurea Linn., but pointed out differences from that species. At that time the relationship of the two mem- bers of this genus, purpurea and manillensis, was not understood. The Ishigaki specimens appear to be typical P. manillensis, though I have had but few skins for comparison. Turnix taigoor (SyKes). Four specimens from Ishigaki, adult @ taken April 25 (wing 77), adult 9 April 25 (wing 84), and two chicks taken April 12. [Eggs, four in a clutch, were taken from March 17 to July 3.] This is the Turniz blakistoni (Swinh.) of Stejneger (Proc. U. S. Nat. Mus., 1886, Vol. IX, p. 635). Dr. Stejneger now agrees with me in the identity of these two forms. Sphenocercus medioximus,? sp. nov. Type. — Mus. Comp. Zoél., No. 37,349. Two adults from Ishigaki, # taken March 9, 9 March 7. Specimens were secured on this island from February 2 to June 5. [Nests containing two eggs each were found on trees at from six to ten feet from the ground, between April 25 and June 2.] 1 Proc. U. S. Nat. Mus., 1891, Vol. XIV. p. 493. 2 Medioximus, middlemost, holding a middle place. 262 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Characters. — Nearest in color to S. permagnus (Stejn.) from the middle group of the Liu Kiu Islands, but much smaller, being little larger than S. formose (Swinh.) of Formosa, Color. — Type, adult g. Forehead yellowish oil green, slightly shaded with chestnut toward crown ; rest of upper parts dark oil green, the feathers of the cervix, sides of head and neck and upper back, pale gray below the green tips, this color showing through a little, giving a hoary cast to these parts ; rump and upper tail coverts a little brighter; primaries slaty black with a percep- tible greenish tinge toward ends, the three outer ones narrowly edged with yel- lowish; secondaries, alula, and middle coverts slaty black somewhat washed with green ; middle coverts and secondaries bordered externally with yellow; rest of wing and scapulars oil green with a slight wash of chestnut on shoul- der; under parts yellowish oil green; middle of belly and striping on flanks yellowish white ; under tail coverts (reaching to end of tail) dark oil green broadly edged with straw yellow ; rectrices above olive green, below slaty black with grayish tips; under surface of wing slaty. “Adult 9, similar to the ¢ but duller in color throughout, and lacking the slight chestnut suffusion on crown and shoulders, and with the grayish tinge of cervix, upper back, and sides of head much less pronounced. Measurements. — Adult ¢, type, wing 193.5; tail 133; tarsus 26.8; exposed culmen 19. Adult 9, topotype, wing 192; tail 129; tarsus 26; exposed cul- men 18.6. The Green Pigeon differs in the islands as follows: S. permagnus is confined to the middle group of the Liu Kius, while S. medioximus is peculiar to the southern group; S. formose belongs further south still, to the island of Formosa. Stejneger’s type of S. permagnus is in the Museum at Tokyo, and I have not seen specimens of the species. In addition to the species here described being intermediate in size between S. permagnus and S. formose, it differs slightly in color from either of the two. In S. medioximus two sets of wing coverts are bordered with yellow, and the male has a decided wash of chestnut on both crown and shoulders. Stejneger especially describes his type as having only one set of coverts “the outer great coverts ” edged with yellow. If the type of S. permagnus be a male, as was supposed, then the chestnut wash on the crown and shoulders of S. medioximus is a distinctive character, and yet again very _ different from the strong coloring of these parts in S. formose. Chalcophaps indica (Lixvy.). Two specimens, # and 9 adults, from Ishigaki. The @ taken March 20, the 9 taken June 10. [Many nests were found, containing two eggs each, usually placed in dead trees at from six to ten feet from the ground.] The two Ishigaki skins differ slightly from two Indian specimens of true C. indica with which I compared them. In the Liu Kiu birds the band on the BANGS: BIRDS FROM THE LIU KIU ISLANDS. 263 lower back between the two gray bands is not coppery bronze, but is dull black, almost without metallic lustre, and the male hasa much greater amount of gray on back and upper neck. A green-winged dove was described by Swinhoe from Formosa as C. formo- sana, but is not recognized as distinct from C. indica by Count Salvadori, in the British Museum Catalogue (Vol. XXI. pp. 514-520). ‘Megascops elegans (Cassy). Two adults from Ishigaki, @ taken March 25, 9 March 23. Specimens were taken from March 1 to June 3. [Eggs, two in a clutch, were taken from holes in trees, seven to fifteen feet from the ground, from May 14 to June 27.] Ninox japonica (Temm. & Scuat.). Three specimens from Ishigaki, adult @ taken April 20, adult 9 April 15, and a half-grown young, no date. These skins agree with Japanese specimens. The wing of the adult ¢ measures 215, of the adult 9 210. Accipiter gularis (Temm. & Scut.). Three specimens, a 9 (?) not in full adult plumage taken June 1, an adult @ March 25, and a downy nestling June 27, all from Ishigaki. Butastur indicus (Gmet.). Two specimens from Ishigaki, neither in full plumage, the # taken June 1, the 9 March 23. Halcyon coromanda rufa (Watace). Two specimens from Ishigaki, adult g and Q, both taken April 25. Speci- mens were secured in Ishigaki and Taketomi from April 5 to June 10. [Eggs, three in a clutch, were collected from June 1 to June 21. The nests were in holes in trees at about ten feet from the ground.] I follow Dr. Stejneger in provisionally referring the Liu Kiu Ruddy Kingfisher to this form. Anthus maculatus Hopes. One female taken in Ishigaki, April 7. Motacilla lugens Kirttt. One adult ¢ in full spring plumage, taken in Ishigaki, June 1. This seems rather a late date for M. lugens to be in the Liu Kiu Islands. = 264 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Hypsipetes pryeri Sreun. Five specimens from Ishigaki, an adult ¢ taken Feb. 29, an adult 9 April 30, and three recently hatched young April 21. [Skins were also obtained in Kabama, and eggs, four in a clutch, were taken from April 2 to June 25.] Merula pallida (Gmet.). ‘ Two adults from Ishigaki, ¢ February 20, 9 May 1. (Many specimens were taken in Ishigaki up to June 20.) Merula chrysolaus (Tremm.). Two specimens from Ishigaki, adult @ May 7, adult 9 February 18. [Skins were collected in Ishigaki between February 18 and June 7. ] Merula obscura (GmeEL ). Two adults from Ishigaki, @ February 22, 9 March 1. [Obtained in Ishi- gaki between February 20 and March 1.] Monticola solitaria (Mitt.). One adult 9, Ishigaki, March 23. Terpsiphone illex,} sp. nov. Tyre. — Mus. Comp. Zodl. No. 37,363. Two specimens from Ishigaki, anadult g¢ April 25, and an adult 9 May 31. [Specimens were taken between April 25 and June 20. Eggs, four in a clutch, between May 12 and June 13.] Characters. — Nearest to T. princeps (Temm.) of China and Japan, but smaller; rectrices narrower and squarer at ends ; wing shorter; primaries very short and decidedly narrower and more pointed at ends; wing formula differ- ent—4th primary longer than 5th (these two equal in T. princeps, or 4th slightly shorter than 5th); feathers of crest in the ¢ all narrower, less rounded; colors much as in 7. princeps, except less white in axillas and lining of wing; feathers of crest in the ¢ steel blue instead of purplish ; sides more heavily washed with brown. The 9 differs from the ? of T. princeps in the same manner as does the @, i. ¢., it is smaller; in having narrower, shorter, more pointed primaries ; narrower rectrices ; crest feathers narrower and bluer, less purplish in color. 1 Tllex, alluring, enticing. BANGS: BIRDS FROM THE LIU KIU ISLANDS. 265 Color.—Adult @, head all round, throat, and jugulum blue black, rather more purplish on throat than on crown; back and scapulars glossy prune purple ; upper tail coverts and tail blue black; wings black edged with purplish brown ; middle of belly and under tail coverts white; sides and flanks heavily washed with dark purplish brown; axillas dull brownish black with white tips ; under primary coverts black ; under wing coverts white streaked with pale brown. Female, crown blue black ; sides of head and cervix dark gray; throat dark gray becoming paler on jugulum; back chestnut, many of the feathers glossy purplish maroon at ends; tail dark purplish brown; wings hair brown edged with hazel, deeply so on secondaries and tertials ; middle of belly and under tail coverts white; sides and flanks washed with purplish brown; lining of wing as in the @, except primary coverts are hair brown instead of black. Measurements. — Adult @, type, wing 88; tail, to end of middle rectrices 246.5, to end of longest other rectrices 113 ; greatest width of outer rectrix, 8.8 ; tarsus 14.4 ; exposed culmen 15.4. Adult 9, topotype, No. 37,364, wing 82 ; tail 80; tarsus 14; exposed cul- men 15.4; width of outer rectrix 9.2. [In adult males of T. princeps the wing ranges from 92 to 94, and the greatest width of the outer rectrix is 11.4. In the adult females the wing measures from 88-90, and the greatest width of the outer rectrix is 12.] Remarks. — This appears to be the first record of a Paradise Flycatcher from the Liu Kiu Islands. Besides being considerably smaller than a T. princeps, it differs noticeably in its short, narrow, pointed primaries and narrow rectrices, and in having the 4th primary longer than the 5th. Like so many of the breeding birds of these islands, it is a well-marked island species. Zanthopygia owstoni,! sp. nov. Type. — Mus. Comp. Zodl., No. 37,367. One male from Ishigaki, June 20. Characters. — Nearest to Z. narcissina of Japan, but wing much shorter, due chiefly to the shortening of the primaries; wing formula different —2nd primary shorter than 6th, 3rd about equal to 5th, 4th longest. In Z. narcissina the 2nd primary is much longer than 6th, 3rd equals 4th, these two longest and longer than 5th. In color the island bird is very different, the back is dark green, not black, the yellow frontal band extends all the way across base of culmen, the throat and breast are clear gamboge yellow, not orange. From Z. zanthopygia (Hay) the species can be distinguished by its yellow eyebrow (white in Z. zanthopygia) and differently marked wing. Color. — Male, apparently fully adult (9? unknown), narrow frontal band, extending directly across base of culmen and thence over eye to the supra- auricular region, gamboge yellow ; pileum, cheeks, back, and scapulars dusky 1 Named in honor of Mr. Alan Owston. 266 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. olive green ; rump bright gamboge yellow; upper tail coverts and tail black ; wings dark hair brown, the lesser coverts dull, dark plumbeous ; a large white wing patch, formed by the white color of the middle and most of the greater coverts ; one or two (on each side) of the longer tertials narrowly edged with whitish for the basal half of the outer web ; throat, jugulum, and breast bright gamboge yellow, becoming yellowish white on belly and under tail coverts ; sides and flanks washed with olive green ; lining of wing and narrow inner margin of wing feathers, below, white. Measurements. — Type @, wing 67 ; tail 45; tarsus 15.8; exposed culmen 10.2; distance from tip of longest secondary to tip of longest primary about 15. Remarks.—In Proce. U.S. Nat. Mus., 1887, Vol. X, pp. 406-407. Stejneger pointed out the structural differences between the Liu Kiu species and Z. narcissina ; he, however, had but one young example of the island species, and on this account refrained from giving it a name. The one skin obtained by Zensaku bears out all the structural characters, and besides shows marked color differences from either Z. narcissina or Z. zanthopygia. The type of Z. owstoni, a male, appears to be in full breeding plumage, and if so, the dark olive green color of the back is unlike any other species, and would alone distinguish the Liu Kiu form. Cisticola brunniceps (Temm. & Scut.). Two adults from Ishigaki, # taken March 7, 9 Junel. The fantail warbler is said to be the most abundant bird in the islands. [It builds its nest in grass a foot or two above the ground. Eggs, as many as seven in a clutch, were taken from March 25 to June 30.] Cettia cantillans (Tem. & Scut.). One adult 9 from Ishigaki, March 5. [Six specimens were taken on Ishigaki between March 5 and April 7.] Cettia cantans (Temm. & Scut.). Two specimens from Ishigaki, ¢ taken March 25, Q April 6. [Specimens were secured between February 18 and April 6.] Hirundo rustica gutturalis (Scor.). Two adults from Ishigaki, ¢ April 4, 9 April 3, 1899. [Four birds in all were obtained on the island between April 2 and April 5.] Pericrocotus tegimae Sresn. A pair of adults from Ishigaki, the 9 taken June 20, the # June 10. These specimens agree exactly with Stejneger’s type. BANGS: BIRDS FROM THE LIU KIU ISLANDS. 267 Lanius bucephalus Temm. & Scut. One adult 9 from Ishigaki, May 10, 1899. I have compared this skin with an extensive series from Japan, and find it identical with mainland birds of the same sex in corresponding plumage. Parus stejnegeri,! sp. nov. Tyre. — Mus. Comp. Zodl., No. 87,392. Three specimens from Ishigaki, adult ¢ February 27, adult 9 June 1, and a nestling June 7. Characters. — Not nearly related to any known species ; general coloration gray-blue, black, and white ; under tail coverts mostly black; outer rectrices With no white, except a very narrow tip on the outer pair; no white patch on nape, a few feathers of this region with partly concealed white spots only noticeable when the feathers are disturbed ; general coloration of nestling greenish and dull yellow, showing the probable affinities of this species to some of the yellow and green titmice, such as P. jerdoni, P. inseparatus, ete., which have black under tail coverts and but little white in the tail. Color. — Adult ¢ type,a large white auricular patch ; rest -of head, throat, jugulum, and neck glossy blue-black ; a few feathers on middle of hind neck with small semi-concealed white spots; back, rump, and upper tail coverts dark plumbeous, slightly paler on lower rump; scapulars and broad edgings to greater and lesser wing coverts plumbeous ; some of the greater coverts tipped with drab-gray, forming a broken and inconspicuous wing bar; rest of wing grayish black, primaries edged with light plumbeous, secondaries with greenish gray, and tertials rather more broadly on outer webs with grayish white; primary coverts greenish gray ; a broad black stripe down middle of under parts, from jugulum to under tail coverts; sides and flanks dull olive gray, much paler and more drabby along edges of central black stripe and below the black of jugulum and sides of neck ; under tail coverts black, slightly edged and tipped with dark plumbeous, one or two of the shortest lateral ones a little marked with white ; rectrices, below blue-black, above, broadly edged on outer webs with dark plumbeous, the central pair mostly of this color, on both webs ; two outer rectrices with very narrow white tips, 2 mm. deep; bend of wing black; under primary coverts black tipped with white; axillas mostly white ; under sides of primaries grayish white on edges of inner webs. Adult 9, topotype, No. 37,393, similar in markings to the male, all the colors duller and lateral under tail coverts more noticeably marked with white. Nestling, topotype, about two-thirds grown, auricular patch olive yellow ; head, back, and throat dusky olive green, darkest on top of head and sides of throat ; a blackish line down middle of belly; sides, flanks, and under tail 1 Named in honor of Dr. Leonhard Stejneger. 268 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. coverts dull olive yellow; wings grayish hair brown, scapulars and lesser coverts dull grayish olive, tips of greater coverts yellowish, forming a wing bar ; primaries and secondaries edged with greenish gray; tail grayish hair brown edged with greenish gray, outer rectrices barely tipped with whitish. Measurements. — Adult @ type, wing 62 ; tail 55.5 ; tarsus 18.2; exposed culmen 11. Adult 9, topotype, No. 37,393, wing 60; tail 50; tarsus 18 ; exposed culmen 10.5. Corvus macrorhynchus levaillantii (Lesson). Four specimens, all from Ishigaki, adult ¢ March 25, adult 9 March 28, and two young from the nest June 10. [Eggs, four in a clutch, were taken April 11 to June 10.] Sturnia pyrrhogenys (Temm. & Scut.). One male from Ishigaki, June 1, 1899. Zosterops loochooensis (Tristram). Two specimens from Ishigaki, adult ¢ March 13, adult 9 April 6. [Abun- dant on Ishigaki and Kuroshima. Skins were taken from February 18 to June 7, and eggs, four in a clutch, April 2 to June 25. ] A careful comparison of these two specimens with numerous examples of Z. simplex and Z. japonica proves the Liu Kiu form to be a distinct island race, in spite of the doubts cast upon it in the latest review of the group.? But as no adequate description of it appears to have been published, I append the following : — Characters. —Nearest to Z. simplex of China, but bill heavier, wing longer ; ofa brighter green color above, and brighter yellow color below; the species differs from Z. japonica in slightly shorter wing and in the color of the sides and flanks, which lack the strong vinous brown of this region in the Japanese species, and also in the primaries, being very short and narrow at tips (a char- acter presented by many of the species of birds peculiar to the Lin Kiu Islands) ; wing formula, Ist primary about equal to 6th, shorter than 5th, 2nd equal to 4th, 3rd longest. Color. — Whole upper parts including margins of wing and tail feathers yellowish oil green, frontal region slightly yellower; wings and tail black (except for the green margins of the feathers); orbital ring silky white; a dusky spot below and in front of eye; chin and throat lemon yellow ; breast and belly soiled whitish, faintly washed with yellowish along median line and with pale écru drab on sides and flanks ; thighs yellowish white in front, dusky oil green behind ; under tail coverts lemon ‘yellows ; bend of wing lemon yellow ; alula black ; Haine of wing and axillas pale yellow ; narrow inner margins to wing feathers below whitish. 1 Finsch, O. Zosteropidae. Das Tierreich, 1901, 16, p. 20. BANGS: BIRDS FROM THE LIU KIU ISLANDS. 269 Measurements. — Adult ¢, No. 37,390, wing 57; tail 39.5 ; tarsus 18; ex- posed culmen 11.2; distance from tip of longest secondary to tip of longest primary 11. Adult 9, No. 37,391, wing 57; tail 40; tarsus 18; exposed culmen 11 ; distance from tip of longest secondary to tip of longest primary 11.5. Emberiza spodocephala Patt. One male, not in full plumage, from Ishigaki, April 8. Passer montanus saturatus Sreun. One adult ¢ from Ishigaki, June 30. This specimen differs from the type of P. saturatus only by slightly paler colors, due to the more abraded condition of its plumage. [The bird was common in the island, and was breeding in the roofs of the houses. Eggs, seven in a clutch, were taken March 20 to June 25.] Coccothraustes coccothraustes japonicus (Temm. & Scut.). One female from Ishigaki, March 7. BULLETIN OF THE MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE, IN CAMBRIDGE. VOL. XXXVII. CAMBRIDGE, MASS., U.S. A. 1900-1901. UNIVERSITY PREss: Joun Witson anp Son, CampripeGe, U.S.A. CONTENTS. PAGE No. 1.— Descriptions of New and Little-known Mrpus# from the Western Atlantic. By Atrrep GoLtpsporoucH Mayer. (6 Plates.) June,1900 1 No. 2.—Some Mepus# from the Tortugas, ‘Florida. By ALrrep GoLps- PaneMcHeMavin, (44 Plates.) July,1900 ....:.. =... . Ji No. 3.— Contributions from the Zodlogical Laboratory of the Museum of Comparative Zoology at Harvard College, under the direction of E. L. Mark. No. 126. The Regenerating Nervous System of LumBricip® and the Centrosome of its Nerve Cells. By Herspert W. Ranp. (8 Plates.) ate IG MEG ek) sk dw ee es ws we oe a BO Bulletin of the Museum of Comparative Zoology AT HARVARD COLLEGE. VoL. XXXVII. No. 1. DESCRIPTIONS OF NEW AND LITTLE-KNOWN MEDUSZ FROM THE WESTERN ATLANTIC. By ALFRED GOLDSBOROUGH MAYER. Wirn S1x PuArEs. CAMBRIDGE, MASS., U.S. A.: PRINTED FOR THE MUSEUM. JUNE, 1900. No. 1. — Descriptions of New and Inttle-known Meduse from the Western Atlantic. By ALFRED GOLDSBOROUGH MAYER. LIST OF SPECIES. SCYPHOMEDUS 4. Bathyluca solaris, nov. gen. et sp. HYDROMEDUSZ. Bougainvillia Gibbsi, nov. sp. Lymnorea borealis, nov. sp. Oceania caroline, nov. sp. Oceania singularis, nov. sp. Octonema gelatinosa, nov. sp. Orchistoma tentaculata, nov. sp. Stomotoca apicata, L. Acassiz. Stomotoca rugosa, nov. sp. = Stomotoca apicata, Fewkes. Syndictyon angulatum, nov. sp. CTENOPHOR. Mnemiopsis McCradyi, nov. sp. Tue Meduse described in the following paper were obtained by the author as assistant to Mr. Alexander Agassiz in collecting new material for a work upon the Medusa-fauna of the Atlantic Coast of North America. The descriptions of Western-Atlantic Meduse herein given will eventually be published also in the new edition of The North American Acalephe now in preparation by A. Agassiz and A. G. Mayer. Eight species are new; of these one is a Scyphomedusa, one a Ctenophore, and six are Hydromedusz. In addition to these there is one Hydromedusa (Stomotoca rugosa) that we have redescribed under a new name. VOL. XXXVII. — NO. 1. 1 2 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. The Scyphomedusa (Bathyluca solaris) is, judging from its structural affinities, a deep-sea type, although the single specimen from which our figures were obtained was found upon the surface of Narragansett Bay, Rhode Island. The Medusz described in this paper were collected at different times at Eastport, Maine; Newport, Rhode Island; Charleston, South Caro- lina; and in the Bahama Islands during visits made to the above localities at the suggestion of Mr. Agassiz. SCYPHOMEDUS &. BATHYLUCA, nov. gen. Bathyluca solaris, nov. gen. et sp. Figs. 1, 2, Plate 1. A single specimen of a new genus of Discomedusa belonging to the family Ephyridze was found in Narragansett Bay, Rhode Island, on July 27, 1896, by R. W. Hall, Esq. The medusa was found floating upon the surface, but as it was very much torn and battered, and as it differs widely from any of the hitherto known pelagic medusz of our coasts, we are inclined to suspect that it may prove to be a deep-sea form, a specimen of which has wandered to the surface. Generic Characters. — Bathyluca, nov. gen. Discomedusze with a simple cru- ciform, central mouth opening, without mouth-arms or palps. There are 16 wide, radial, gastro-vascular pouches (8 ocular and 8 tentacular). There is no ring canal. There are 8 marginal sense-organs and 16 marginal ten- tacles. There are 4 gonads in the oral floor of the disk, and there are 4 sub-genital pits. Specific Characters. — The umbrella is flat, and the gelatinous substance is quite thick. It is 45 mm. in diameter, and about 10 mm. in height. The aboral surface of the umbrella is sprinkled over with small clusters of nemato- cysts. There are 8 marginal sense-organs that are deeply sunken within small niches between the lappets. The entoderm of these sense-organs contains no pigment, but instead there are small white granules (Fignre 2). There are 24 marginal lappets and 16 long hollow tentacles. The mouth opening is cruciform in shape, and there appear to be no mouth-arms or palps. We may, however, be mistaken in regard to this, for our specimen was much corn and battered, and it is possible that the palps may have disappeared. There are 4 wide sub-genital pits. The gonads are found in the entoderm of the lower floor of the gastro-vascular cavity, and their position is marked by 4 horseshoe-shaped ridges upon the lower floor of the sub-umbrella. There are a number of long gastric cirri that arise from the regions of the gonads MAYER: MEDUS4 FROM THE WESTERN ATLANTIC. 3 and project slightly beyond the mouth opening. The stomach is large, its diameter being about 4 that of the umbrella itself. Sixteen wide, simple radial pouches extend outward from the stomach cavity into the peripheral regions of the umbrella. Hight of these pouches go to the marginal sense- organs, and 8 to the tentacles which are hollow throughout almost their entire length. There are 8 radial bands of muscle fibres in the ex-umbrella. These go to the marginal sense-organs. The gelatinous substance of the disk is translucent but slightly bluish in color. The clusters of nematocysts over the aboral surface are dull yellowish brown, and the tentacles are slightly green in color. Single specimen, Narragansett Bay, Rhode Island. HYDROMEDUS A. STOMOTOCA, L. Aeassiz, 1862. Stomotoca apicata, L. Acassiz. Fig. 3 ¢, Fig. 4 9, Plate 2. Saphenia apicata, McCrady, J., 1857, Gymn. Charleston Harbor, p. 27, Pl. VIIL. Figs. 2, 3. Male Stomotoca apicata, Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV. p. 347. Stomotoca apicata, Agassiz, A., 1865, North Amer. Acal., p. 168. Dinematella cavosa, Fewkes, J. W., 1881, Bull. Mus. Comp. Zoél. Har- vard Coll., Vol. VIII. p. 151, Pl. Il. Figs. 2,3; Pl. IV. Fig. 3. Female | cic cavosa, Fewkes, J. W., 1884, Amer. Nat., Vol. XIX. p. 195, l Fig. Stomotoca apicata, L. Agassiz, is distinguished by the fact that the entoderm of the proboscis in the male is emerald green, or straw-colored; and in the female dull ochre. Also the tentacle bulbs in the male are purple, and in the female dull ochre. This species has been confounded by Brooks, 1883, and Fewkes, 1881, with another form in which the entoderm of the proboscis and tentacle bulbs is brick-red in both sexes. For this brick-red form we propose the name Stomotoca rugosa. Specific Characters. — Stomotoca apicata. In the adult medusa the bell is about 4 mm. high and 2 mm. broad. It is provided with a prominent apical projection that is solid in the males, but usually hollow in the females, the gastro-vascular space leading upward into it. There are two long tentacles with large, hollow basal bulbs. In addition to the two long tentacles there are usually 6 small rudimentary tentacle bulbs upon the bell margin. The pro- boscis is flask-shaped, there is no peduncle, and the 4 lips are curved slightly upward. The ectoderm of the upper portion of the proboscis, under the 4 radial tubes, is thrown into folds or convolutions, and it is in this region that 4 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. one finds the gonads. There are 4 broad radial tubes and a broad circular vessel with somewhat jagged outlines. The velum is well developed. The color of the proboscis in the male varies from intense green to dull ochre- yellow, or cream-color; and the basal bulbs of the tentacles vary from faint to deep purple. In the females, the proboscis and tentacle bulbs are usually dull ochre-yellow, or cream-color, but in some few individuals the proboscis is faintly straw-colored, and the tentacle bulbs faint purple. In the female the apical projection of the bell is hollow, while in the male it is usually solid. Common at Newport, Rhode Island, from July 15-September. Rare at Charleston, South Carolina. The young medusa resembles the adult excepting that the apical projection to the bell is wanting, or is but little developed. There are 2 tentacles and 2 rudimentary tentacle bulbs. The sexual color difference is seen in the young- est medusz we have observed. The hydroid stock is unknown. Stomotoca rugosa, nov. sp. Fig. 5, Plate 2. Stomotoca apicata, Fewkes, J. W., 1881, Bull. Mus. Comp. Zool., Vol. VIIL. p. 152, Pi bics, 1459" Amphinema apicatum, Brooks, W. K., 1883, Studies Biol. Lab. Johns Hopkins Univ., Vol. II. p. 473. The bell is 5 mm. high and 3 mm. broad; it bears an apical projection which in some specimens is long and slender, and in others is short and blunt. The substance of this projection is solid throughout. There are 2 long, well- developed tentacles and 14 small rudimentary ones. The basal bulbs of the long tentacles are large and hollow. When fully stretched, the long tentacles attain a length of 4-6 times the bell height. The velum is well developed. There are four broad radial tubes, and a broad circular vessel with jagged outlines. The proboscis is flask-shaped, the lips being flanged and quite prominent. The sexual products are found in the ectoderm of the upper portion of the proboscis where the outer surface is folded into a complex series of ridges. The bell is transparent, and the entoderm of the tentacle bulbs and of the proboscis is brick-red. In some individuals the 4 radial tubes and the circular vessel are faint red. There is a well-marked southern variety of this species, found at the Tortu- gas, Florida, in which the proboscis and the tentacle bulbs are brick-red streaked with black. In some individuals, indeed, the proboscis and ten- tacle bulbs are coal-black. Brooks, 1883, has described the hydroid and young medusa of this species from Beaufort, North Carolina. According to him, the hydroid stock is a Perigonimus very much like P. minutus, Allman, 1871, p. 324, Plate XI. Figures 4-6. This medusa is common at Newport, Rhode Island, and is also found at Charleston, South Carolina, It is rare at the Tortugas, Florida. MAYER: MEDUSA FROM THE WESTERN ATLANTIC. 5 SYNDICTYON, A. Aeassiz, 1862. Syndictyon angulatum, nov. sp. Figs. 6-8, Plate 3. Specific Characters. — The bell is almost square in cross-section and is not quite as broad as it is high. The bell height in the specimens found by us was about 2.5 mm. There are 4 stiff tentacles that are about three-fourths as long as the bell height. The distal halves of these tentacles are conical in shape, and are covered thickly with clusters of nettle cells.. The basal bulbs of the tentacles are large and swollen, and contain each a single well-developed ectodermal ocellus. This ocellus is formed by a cup-shaped invagination of ectodermal cells that are deeply stained with dark-brown pigment granules. It is probable that this structure constitutes a very primitive udoscopic eye. The velum is small. There are 4 narrow, straight, radial tubes and a slender circular vessel. The proboscis is spindle-shaped, and the mouth is a simple circular orifice. The gonads are situated within the ectoderm of the proboscis. The entoderm of the proboscis and of the tentacle bulbs varies from turquoise to blue-green in different specimens. Several specimens of this medusa were found off Turks Islands, Bahamas, January 20, 1893. BOUGAINVILLIA, Lesson, 1836. Bougainvillia Gibbsi,! nov. sp. Figs. 14, 15, Plate 4. Specific Characters. — Adult medusa ; Figure 14. The bell is about 4 mm. in height and 3.8mm. in diameter. The gelatinous substance is very thick, so that the bell cavity is only about one half as deep as the height of the animal. There are 4 clusters of marginal tentacles which arise from 4 large bulbous swellings, situated at the bases of the 4 radial canals. Each bulbous swelling gives rise to 4 or 5 long slender tentacles. There is a single dark-brown ocellus at the base of each tentacle upon the centripetal (lower) side. The velum is small. There are 4 straight, narrow, radial canals. The proboscis is wide and cruciform in cross-section, and the radial canals arise from the 4 corners of the cross. The proboscis is short and does not extend quite one half the distance from the inner apex of the bell cavity to the velar opening. The mouth is situated at the extremity of a short tubular neck, and there are no prominent lips. Four radially situated oral tentacles arise from the sides of the neck of the proboscis. Each one of these branches dichotomously about twice. The gonads are developed upon the sides of the stomach, and 1 Named for Mrs. Theodore K. Gibbs. 6 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. in the female the ova are large and prominent. The proboscis is pearl- colored, or of a delicate green. The entodermal cores of the tentacle bulbs are red surrounded by a delicate yellow-green. The supporting lamella of the bell often displays a faint greenish tinge. Young Medusa. — In the young medusa there are but 8 tentacles, 2 from each tentacle bulb. The bell is a little higher than a hemisphere and the gelatinous substance is not very thick, being of about uniform thickness every- where instead of being very thick at the aboral pole, as in the adult. The proboscis is short and quadratic, and there are 4 short, unbranched, knob- shaped oral tentacles. When the medusa is about 3 mm. in height, the bell is still hemispherical. The proboscis is wide, shallow, and quadratic, and the oral tentacles branch once dichotomously. About 4-5 marginal tentacles arise from each tentacle bulb. This medusa is found in Newport Harbor, Rhode Island, from July until October. This species is distinguished from Margelic carolinensis, L. Agassiz, by the greater height and less width of its bell. Also in M. carolinensis the pro- boscis is long and slender, while in B. Gibbsi it is short, wide, and cruciform in cross-section. The proboscis of M. carolinensis is widest at about the middle point of its length, while that of B. Gibbsi is widest at its proxi- mal base. LYMNORBEA, Péron and Lesvecr, 1809. Lymnorea borealis, nov. sp. Figs. 16-18, Plate 5. Specific characters. — The bell is 3 mm. in height. The bell walls are thin, and there is a slight apical projection. There are 32 well-developed marginal tentacles with large basal bulbs. These tentacles are about } as long as the bell height, and are curled slightly upward. They are not very flexible. The velum is well developed. There are 4 straight, narrow radial tubes. The proboscis is pyriform and the mouth is surrounded by 4 short, dichotomously branching, oral tentacles. Each of these oral tentacles branches 2 times, thus giving rise to 4 tentacle tips (see Figure 18). These tips are short and knob-like and are covered with long slender nematocyst capsules borne upon thread-like filaments (see Figure 17). The gonads oceupy 4 radially situated, longitudinal swellings upon the proboscis. The entoderm of the proboscis, and of the bulbs of the marginal tentacles, is red. Three specimens, all of them being males, were found in Eastport Harbor, Maine, on September 19, 1898. MAYER: MEDUS#Z FROM THE WESTERN ATLANTIC. 7 OCEANIA, Péron and Lesvevr, 1809. Oceania caroline, nov. sp. Figs. 9-11, Plates 3, 4. Specific Characters. — The bell is not quite a hemisphere, and is 14 mm. in diameter. The cavity of the bell is shallow, so that the gelatinous substance is quite thick. There are 16 well-developed marginal tentacles with large, hollow basal bulbs. These are only about half as long as the bell diameter, but as they are usually carried coiled in a close helix they appear much shorter. In addition to these well-developed tentacles there are 48 small rudimentary tentacle bulbs that probably never develop into tentacles. There are 64 otocysts, 4 between each adjacent pair of large tentacles (see Figure 11). Each otocyst contains 2 spherical otoliths. The velum is well developed. There are 4 narrow, straight, radial canals. The mature proboscis (Figure 10) is flask-shaped, and there are 4 simple curved lips. The gonads are de- veloped upon the radial tubes at about one quarter the distance from the circular vessel to the proboscis. In the female the ova are very conspicuous. The entoderm of the tentacle bulbs and proboscis and of the radial tubes in the region of the gonads is bright yellow-green. This species was extremely abundant in Charleston Harbor in the early part of September, 1897, and in June, 1898. Oceania singularis, nov. sp. Figs. 12, 13, Plate 4. Specific Characters. — The bell is 2 mm. in diameter and the sides are quite straight and sloping. Near the apex of the bell there is a sharp constricticn, above which there is a lens-shaped apical projection. There are 16 well-de- veloped marginal tentacles with large, hollow, conical-shaped basal bulbs. The lashes of the tentacles are short and are covered with nematocystic cells. In addition to the 16 functional tentacles there are 16 intermediate rudimentary ones. There are 32 otocysts, each containing a single highly refractive spheri- cal otolith. There are 4 straight radial tubes. The proboscis is quadrangular in cross-section, and there are 4 simple lips. The 4 gonads are developed upon the 4 radial canals near the base of the proboscis. The entoderm of the proxi- mal part of each tentacle bulb is turquoise-green, and the distal part is brownish-red. The entoderm of the proboscis and of the radial tubes in the neighborhood of the gonads is of a delicate turquoise tinge. A single specimen of this medusa was found in Newport Harbor, Rhode Island, on August 22, 1896. 8 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. OCTONEMA, Hacrcxet, 1879. Octonema gelatinosa, nov. sp. ; Figs. 20, 21, Plate 6. Specific Characters. — Young medusa? The bell is 3.5 mm. in diameter and somewhat flatter than a hemisphere. The gelatinous substance is quite thick. In the single specimen examined there were 4 tentacles with long hollow basal bulbs. These tentacles were about 2 times as long as the bell diameter. Within the entoderm of the inner side of each tentacle bulb there was a single dark-colored pigment spot. In addition to these long tentacles there were 12 rudimentary tentacle bulbs upon the bell margin. It is possible that these might have in time developed tentacles ; in the specimen observed by us, however, they were very small and apparently rudimentary. A dark- colored pigment spot was found in the entoderm of each of these tentacle bulbs. There were 8 marginal clubs, 2 in each quadrant. A dark-brown entodermal pigment spot was situated at the base of each (see Figure 21). The velum was well developed. There were 4 straight radial canals upon the upper regions of which the gonads were situated. The proboscis was a simple tube with 4 simple lips. The color of the entoderm of the 4 large tentacle bulbs, and of the radial tubes in the region of the gonads was green. A single specimen was found in Charleston Harbor, South Carolina, September 14, 1897. ORCHISTOMA, Haercket, 1879. Orchistoma tentaculata, nov. sp. Fig. 19, Plate 5. Specific Characters. — Young medusa: The bell was 6 mm. in height. The sides near the margin were slightly flanged outward. The gelatinous substance of the upper portion of the bell was very thick,so that the concavity was shallow. There were thirty-two marginal tentacles in various stages of devel- opment, the longest being about 1.5 times as long as the bell height. The tenta- cles possessed long, hollow basal bulbs. There were no marginal sense-organs. There were sixteen functional radial tubes, and sixteen others in process of development. The radial tubes were straight, and there were no traces of gonads upon them. The velum was well developed. The proboscis was flat and shallow, and there were 8 lips. The entoderm of the basal bulbs of the tentacles was of a delicate green. Only one immature specimen of this medusa has ever been seen ; it was found at Newport, Rhode Island, August 18, 1896. The genus is closely related to Melicertum. MAYER: MEDUSA FROM THE WESTERN ATLANTIC. 9 CTENOPHORZ. MNEMIOPSIS, L. Acassiz, 1860. Mnemiopsis McCradyi, nov. sp. Figs. 22, 23, Plate 6. Specific Characters. — Mnemiopsis McCradyi: This species is closely allied to Mnemiopsis Leidyi, A. Agassiz, but differs from it chiefly in the much greater complexity of the ramifications of the chymiferous tubes within the lappets ; and also in the very decided amber color of the gelatinous substance of the animal. It is also remarkable that in this species the gelatinous sub- stance of the body is of so tough a nature that the creature may be removed from the water by hand without suffering injury. Indeed, we know of no Cteno- phore that is as resistant as this species. The animal is 100 mm. in length, our figures being natural size. There are eight longitudinal rows of ciliated plates. Four of these rows, that extend down the lateral lappets, are about twice as long as are the four others that lead from the apex to the auricles. The body is markedly compressed, the broad lateral axis, extending through the lappets, being about twice as great as the auricular axis. (Compare Fig- ures 22 and 23.) The lateral lappets are about as Jong as the remaining por- tion of the body, and are similar in shape and size to those of M. Leidyi, and much longer than in M. Gardeni. (See A. Agassiz, 1865 ; North American Acalephe, Figures 20, 21 and 22, 23.) The apical sense-organ is found at the bottom of a deep cleft at the aboral pole of the body, and is similar in structure to that of M. Leidyi. The chymiferous tubes that wind through the lateral lappets are of a decided purple color and their ramifications are very complex. The mature ova are similar in appearance to those of M. Leidyi. A single perfect specimen of this species was found in Charleston Harbor, South Carolina, September 15, 1897. 1 This species is named in honor of Professor John McCrady in recognition of his important researches upon the medusz of Charleston Harbor. MAYER. — Western Atlantic Meduse. PLATE YE: Fig. 1. Bathyluca solaris, nov. gen. et sp. Oral view of the medusa. Fig. 2. Bathyluca solaris. Oral view of one of the marginal sense-organs. mi ’ i ee a . B Metsel lith Boston. PLATE STERN ATLANTIC MEDUSAE. f Ay ui is » @ iP a hv, ‘Sie ‘ % as - s * 2 De ieee > » a NS ae aie 7 « © 4 - <= “ ‘ j pl be ye p ; hl Mayer. — Western Atlantic Meduse. PLATE 2. Fig. 5. Stomotoca apicata, L. Agassiz. Male medusa. Fig. 4. Stomotoca apicata. Female medusa. Fig. 5. Stomotoca rugosa, Mayer = Stomotoca apicata, Fewkes, 1881. TERN ATLANTIC MEDUSAE. PLATE 2. B Meisel lith Sesten. Mayer. — Western Atlantic Medusz. PLATE 3. Syndictyon angulatum, nov. sp. Side view of medusa. Syndictyon angulatum. Side view of one of the tentacle bulbs, showing the ocellus. Syndictyon angulatum. Surface view of tentacle bulb. Oceania caroline, nov. sp. N ATLANTIC MEDUSAE. PLATE. 3. ’ B Meisel lith Bosten. a a ee ae - 1” oh i “Vv © ope es - . 7 y «< . ‘ ae . - f = % Mayer. — Western Atlantic Medusz, Fig. 10. Fig. 11. Fig. 12. Fig. 13. Fig. 14. Fig. 15. PLATE 4. Oceania caroline, nov. sp. Side view of proboscis and radial canal. Oceania caroline, nov. sp. View of bell margin. Oceania singularis, nov. sp. Oceania singularis. View of bell margin. Bougainvillia Gibbsi, nov. sp. Mature medusa. Bougainvillia Gibbsi, nov. sp. Young medusa. WESTERN ATLANTIC MEDUSAE. . PLaTE 4, B Meisel lith Boston. ee oS eee a ie Mayer. — Western Atlantic Medus. Fig. 16. Fig. 17. Fig. 18. Fig. 19. PLATE 5. . 3 ' Lymnorea borealis, nov. sp. ° Lymnorea borealis, nov. sp. View of nematocyst capsules upon the | tentacles. > oa Lymnorea borealis. Side view of proboscis, showing the oral tentacle aaa Orchistoma tentaculata, nev. sp. = Prars | &; fair qenshattnorene ear Sa Urls Si = 6 ie Tay ¢ 7 oF ' - Pete nome ae A isi. \ B Meisel lith Boston, al Mayer. — Western Atlantic Meduse. ‘ PLATE 6. See co. \ Lire aa Fig. 20. Octonema gelatinosa, nov. sp. i ® Fig. 21. Octonema gelatinosa, nov. sp. Marginal sense club. . =) ton siele eee Niobia dendrotentacula, nov.'ger.etsp. . . - . . | « . +. 6.50 3 ee Obeliasisps <2 3% SR apt te, ad ry a eel) SOY kee, eg an Oceania eae nov. ap: beet, wil ikea pe onl om Go ere eo Bie rn Oceania :gelatinosa, Dovispie ef se a YO Ae A, 2 Oceania globosa, nov. sp. . Serer Oceania McCradyi = penthen MeCradya Brooks ee Oceania magnifica, nov. sp... Sees ae win & © on Oo ee Pandea violacea, Agassiz and Mayer eee rome hee Ce SL Phortis lactea, nov. sp. . sf 13, Style oe ees Phortis pyramidalis = Feutiner ay eondake L. Ronee 3 a, ne ae cet Pseudoclytia, pentata, Dov. genetisp. ~ <=... - = = 3 Linerges mercurius, Haeckel Nausithoé punctata, Kolliker MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 15 III. SIPHONOPHOR. PAGE Abyla pentagona, Eschscholtz © - - - - + se es are Bee eee a Peiemacuboides, CHO a. 4 tes 77 Abyla quincunx, @lrin? wee Ride t sak AU Ae BE ab 2 at dio, } Aglaisma quincunx = a pisisnioides Gameuhe! Chun Sols ha de Otel tail tseh EA abaete ue Re) Agalma Pourtalesii, Agassiz and Maver Woes Sbder wh Sea ee er me Agalma virida, nov. sp. - - BR ASC Net Load em NOt OPT MCS Bag ati t: Mrcalea =. Chunia capillaria, nov. gen. et = PAE a Sore cette OF cee Lalas "oo ey Ns COS 2 a Cc 74 Eudoxia campanula, Leuckart. © - - - + + ss ss tt ett 74 { Diphyopsis campanulifera, Gili be) ed Rey eee 0 a ed Fe Ersea Lessonii, Chun. . - Pe eRe en a trees ad, otge oP og a § Diphy opsis ESCUELA) ane GSR A 76 | Erszea hispaniana, nov. sp. - ey ae ha ce) er me 1 Diphyopsis picta = Doramasia picta, ‘Chun SAR ra Psi se i hacen PUREE DING SS capt.) cou yh baa nite AN ee CP RP Ft 75 Deere tpelagica, BORG, 6. ee ee SR tt 73 Sueemiimanesna, Tecson - 2 st tt te 72 Rhizophysa clavigera, Chun - © - - - 2 / se ee et tt 72 Rhizophysa Eysenhardtii, Gegenbaur - © - - 2 2 ss es tts 72 Rhizophysa Murrayana,Chun. . . - - - sss ss Sr tt te 72 ; Sphzronectes praciis, Faeckels: 602 te, dye ce Fees wl Fe ae 73 Diplophysa miprmata Geeenbaur 60. pty ls roy a esenier et ey os) fa,” 74 = Sy ar 2S ST ee TC Cc 71 IV. CTENOPHOR. Beroé Clarkii = Idyiopsis Clarkii, L. ais Dae Sega s MRT PAR eee Bolina vitrea, L. Agassiz. . . : De ys) olen ees ss) | oratorio tel Eucharis multicornis ? Tieahacholts ARTA goa Rm: ee se ara aaa Oe eee Hormiphora plumosa ? eRe sa fe ae et cas tint ee OE peel? Oo owen Oe Ocyroé Pememeiea Tans ee Re ad ees nl le 81 INTRODUCTION. Tur meduse described in the following paper were obtained by the author while assistant to Mr. Alexander Agassiz in collecting for a work upon the Medusa fauna of the Atlantic Coast of North America. Three expeditions were made, for Mr. Agassiz, to the Tortugas, Florida, extending from June 10-22, 1897; June 25- August 19, 1898; and May 14-July 4, 1899. The manuscript has been submitted to him, and the descriptions herein given will ultimately be published also in the new edition of “North American Acalephe ” now in preparation by A. Agassiz and A. G. Mayer. We wish to avail ourselves of this opportunity to express our appre- ciation of the cordiality and kindness of George R. Billbury, Esq., head 16 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. keeper of the lighthouse at Loggerhead Key, Tortugas, to whose in- telligent and painstaking co-operation we owe much that may be of value in the following paper. We also wish to thank Major J. E. Sawyer, U. S. A., to whose permission we were indebted for the use of the government steamer, “George W. Childs,” for transportation to and from between Key West and the Tortugas. We are also indebted for a like service to the officers of the Union Bridge Company in allowing the use of their steamer ‘‘ Ambrosio Bolivar.” The Tortugas occupy what is probably the most favorable situation from which to study the pelagic life of the Tropical Atlantic. They lie - upon the northern edge of the deep channel of the Gulf Stream as it issues from the Gulf of Mexico. Pure, deep ocean water- surrounds them, and there are none of the shallow mud-flats that render the shore waters of Florida so turbulent, at times, that many of the more delicate pelagic animals are killed. As is well known, the Gulf Stream pours outward from the Gulf of Mexico through the Straits of Florida. The Gulf Stream does not occupy the whole cross-section of the strait, how- ever, but according to the researches of Lieutenant, now Commander, J. E. Pillsbury (Report U. S. Coast and Geodetic Survey, 1885-87), it flows nearer to the Cuban coast than to the line of the Florida Keys. The northern limit of this great stream lies at least 28 miles south of Rebecca Shoal, the average edge being about 6 miles farther south, or 34 miles south of Rebecca Shoal (see U. S. Coast Survey Report, 1887, pp- 174, 175, Illustration 42). The currents in the immediate vicinity of the Tortugas are extremely variable and are greatly under the influence of the tides and winds, while the tides themselves are small and easily influenced by extraneous circumstances. In the passage between Rebecca Shoal and the Tortu- gas the current sets practically north with the flood tide and south with the ebb. About five miles west of Loggerhead Key the southerly set of the ebb tide is stronger than the northerly current induced by the flood. There can be no doubt that the prevailing winds play an im- portant part in setting up local currents in the immediate vicinity of the Tortugas, The prevailing E.-S.E. winds of the summer months cause a decided westerly surface drift, and this is evidenced by the fact that dur- ing this period sand is washed away from the eastern shore of Logger- head Key and spread out into long cuspate forelands! which extend from 1 “ Cuspate foreland” is a term used by F. P. Gulliver (1896 ; Bull. Geol. Soc. America, Vol. VII.) to denote a sandy, projecting point of land which has cuspate outlines, and is formed by the agency of currents. MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. LZ both the north and south ends of the island in a westerly direction. The island thus assumes, roughly, the form of a crescent with its horns pointing westward. The north winds that occur during the winter months annually destroy these crescentic horns, but they are annually replaced by the summer breezes. Although the northern edge of the current of the Gulf Stream prob- ably never impinges against the Tortugas, a fresh south breeze is suffi- cient to drive its surface waters, unaccompanied by the current,! upon the islands, and under these conditions vast quantities of gulf-weed, and large numbers of Physalia, Velella, and other pelagic animals are cast up upon their shores. It is well known that the Gulf Stream bears along upon its surface vast numbers of floating animals that are drawn into it by winds and currents from the adjacent tropical regions of the Atlantic, and thus it comes about that pelagic animals from all over the Gulf of Mexico and West Indies are drifted past the Tortugas. The temperature of the surface waters in the immediate vicinity of the Tortugas is remarkably high, being about 74°-77° F. in winter, and 80°-86° F. in summer, the average for the whole year being about 78° F. It is probably owing to this high temperature, and also to the great purity of the ocean water, that marine animals may be maintained alive in aquaria with remarkable success at the Tortugas; for the tem- perature of the laboratory is almost sure to be lower than that of the sea, and thus the animals in the aquaria are refreshed and thrive well. CoMPARISON OF THE TorTUGAS FAUNA WITH THAT OF THE SouTHERN Coast or New ENGLAND. Ninety species of Acalephs have been found at the Tortugas. Of these, 62 are Hydromeduse, 16 Siphonophore, 7 Scyphomedusz, and 5 Ctenophore. Of these, 39 species are new to science, 33 being Hydromedusz, 3 Siphonophore, 1 Hydroid, and 2 Scyphomeduse. The Acalephian fauna of the Tortugas is strictly tropical, and is totally different from that of the eastern coast of New England north of Cape Cod. A number of characteristic Tortugas forms are, however, blown northward every summer, and are thus found in considerable numbers upon the southern coast of New England, where they have been found in Newport Harbor and in Buzzard’s Bay. Only three Tortugas species have, however, succeeded in establishing themselves 1 See Lieutenant (now Commander) J. E. Pillsbury, 1886, Report of U. S. Coast and Geodetic Survey, Appendix No. 11, p. 287. 18 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. in Buzzard’s Bay and Newport Harbor; these are: Turritopsis nutri- cula, Margelis carolinensis, and Stomotoca rugosa. But these northern specimens of the two latter forms display distinct and constant color differences which distinguish them from their near relatives in the Tortugas, and probably entitle them to rank as varieties one of the other. In addition to these three Hydromedusz, there is one Scy- phomedusa, Dactylometra quinquecirra, that is established in Tampa Bay, Florida, and also in the bays and estuaries of the southern coast of New England. It has not yet been found at the Tortugas, but, judging from its range of distribution, it probably will be discovered there. There are a number of other characteristic Tortugas Acalephs that may be classed as occasional visitors to the southern coast of New England, upon which they are drifted by the agency of the prevailing S.-S.W. winds of the summer months. None of these appear to suc- ceed in establishing themselves permanently upon the New England coast. Among these Hydromeduse may be mentioned, Eutima mira, /Xquorea floridana, Glossocodon tenuirostris, and Liriope scutigera ; and among the Siphonophore, Physalia pelagica, Velella mutica, Porpita Linneana, Diphyes bipartita, Eudoxia campanula, Ersza Lessonii, Diphyopsis campanulifera, and Diplophysa inermis. No doubt further researches will increase this list of tropical Acalephs that are drifted far from their southern habitat and slowly perish in the colder waters of the north. It is interesting to notice that the Acalephian fauna of Charleston Harbor, South Carolina, in latitude 32°, 20’, is very different from that of the Tortugas, and may be said to be subtropical; for it is intermediate in character between the fauna of the Tortugas and that of the southern coast of New England. For example, the following 13 Acalephs are established both at Charleston, South Carolina, and on the southern coast of New England: Dactylometra quinquecirra, Cyanea versicolor, Eucheilota duodecimalis, Epenthesis bicophora, Oceania languida, Willia ornata, Gemmaria gemmosa, Pennaria tiarella, Stomotoca rugosa, Stomotoca apicata, Turritopsis nutricula, Margelis carolinensis, and Nemopsis Bachei; and the following 17 Acalephs are found both at Charleston and the Tortugas: Dactylometra quinque- cirra? Beroé Clarkii, Bolina vitrea, Margelis carolinensis, Stomotoca rugosa, Gemmaria gemmosa, Turritopsis nutricula, Halitiara formosa, AMquorea floridana, Eutima mira, Eutimalphes coerulea, Epenthesis fol- leata, Eucheilota ventricularis, Steenstrupia gracilis, Liriope scutigera, MAYER: MEDUSZ FROM THE TORTUGAS, FLORIDA. 19 Glossocodon tenuirostris, and Dyscannota gemmifera. In addition to these there are a few Acalephs such as Stomolophus meleagris that are strictly subtropical, having been found neither at the Tortugas nor upon the southern coast of New England, but which are abundant at Charleston. It is important to observe, also, that the Acalephian fauna of the Bermudas, like that of Charleston, is distinctly intermediate between the fauna of the Tortugas and that of the southern coast of New England. Of the 30 species described from the Bermuda Islands by Fewkes (1883; Bull. Mus. Comp. Zool. at Harvard Coll., Vol. X1.), 9 are established at Newport, Rhode Island; and 16 at the Tortugas. Not a single species of acaleph known from the Tortugas has been found established upon the eastern coast of New England north of Cape Cod. The fauna of the eastern coast of New England is, however, closely related to that of the northern coast of Europe (see Browne, 1895, Trans. Liverpool Biol. Soc., 96; Proc. Zoél. Soc., London; Hart- laub, 1897; Helgolands Medusen, etc.). To summarize, then, we have at the Tortugas a tropical fauna that gradually disappears, and is replaced by other forms, as we go north- ward along the coast of the United States. Only three species of the Tortugas fauna are established upon the southern coast of New England, and not one extends north of Cape Cod, Massachusetts. It appears that the great majority of the forms established at the Tortugas are incapable of surviving in the colder waters of the north, although individuals are annually driven far to the northward of their natural habitat by the agency of the Gulf Stream, and the prevailing S.-S.W. winds of the summer season. CoMPARISON OF THE TorTUGAS FAUNA WITH THAT OF THE TROPICAL ATLANTIC. Very instructive facts are brought to light when we compare the Acalephian fauna of the Tortugas with that of the warm zone of the Atlantic Ocean. By the term “warm zone” we include all that region of the Atlantic lying between 30° N. Lat. and 10°S. Lat., and extend- ing from the coast of Africa to the American shores. This “ warm zone ” includes the Canary and Cape Verde Islands, the Bahamas and West Indies, the Guinea Stream, the North and South Equatorial Currents, and the warmer parts of the Gulf Stream. At the present time about 130 species of Hydromeduse are known to inhabit this “ warm zone.” 20 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Haeckel, 1879, describes 30 species from the Canary Islands, 10 from the coast of Africa and Cape Verde Islands, and 10 from the “Tropical Atlantic.” Maas, 1893, in his account of the Hydro- medusze of the Plankton Expedition, enumerates about 21 additional species; and 57 others have been made known by L. Agassiz, Brooks, Fewkes, and Mayer from the Bahamas, Florida Reefs, and Tortugas. The Hydromedusan fauna of the Tortugas is so closely related to that of the Florida Reefs and the Bahama Islands, that they may be said to be practically identical; and we will therefore speak of it hereafter as the “ Bahama-Tortugas”’ fauna. When we come to compare the Hydromedusan fauna of the Bahama- Tortugas with that of the remaining portion of the ‘‘ warm zone,” ex- clusive of the West Indies, we are met with the remarkable fact that only 7 species are known to be common to both the Bahama-Tortugas region and the great remaining region of the “warm zone.” Thus only 5 Hydromeduse have been found in both the Canary Islands and Bahama-Tortugas region. These are Adginella dissonema, Aglaura hemistoma, Aglaura hemistoma var. Nausicaa, Staurodiscus tetrastaurus, and Laodicea ulothrix. Two other Hydromedusz, Glossocodon tenui- rostris and Liriope scutigera, are found in the midst of the ocean between the Canary Islands and the West Indies. It will be noticed that 5 out of these 7 forms that are common to both the eastern and western halves of the “warm zone” are Trachylina, or forms that develop through a free-swimming planula and pelagic actinula stage. The two others, Laodicea ulothrix and Staurodiscus tetrastaurus, belong to the Leptolinidee and probably develop through a sessile hydroid stage with alternation of generations. In 1893 it was shown by Maas in “ Die Craspedoten Medusen der Plankton-Expedition,” and in Natural Science, Vol. II. pp. 92-99, that the great majority of the Hydromedusze found in the midst of the Atlantic, far from land, belong to the Trachylina, and the few Leptolina discovered always show a relation to some neighboring coast. As is well known, it was the avowed object of Hensen’s Plankton Expedition of 1889 to study the organic life of the high seas as free from the influence of * coasts as possible. This expedition entered the region that we have designated the “ warm zone” on August 20, and left it on October 20, 1889. During these two months the expedition remained for by far the greater part of the time upon the high seas, approaching land only at the Cape Verde Islands, Ascension, Fernando Noronha, and the mouth of the Amazon. As has been shown by Maas, 1893, the Hydromeduse MAYER: MEDUSZ FROM THE TORTUGAS, FLORIDA. 21 found in this region consisted almost entirely of forms of Trachylina, composed of Trachynemide, Aglauride, and especially Geryonide. (See Craspedoten Medusen der Plankton Expedition, 1893, Taf. ma VII.) The facts then appear to be that we have at the eastern extremity of our ‘warm zone,” or in that region adjacent to the coast of Africa and in the neighborhood of the Canary Islands, a Hydromedusan fauna com- posed of both Trachylina and Leptolina, and the species which compose this fauna show a distinct relationship with Mediterranean forms. In the midst of the “warm zone,” midway between the Canary Islands and the West Indies, the fauna is composed almost entirely of forms of Trachylina that are pelagic species par excellence, and are distributed widely over the high seas, and also reach the coasts of Africa and America. In the Bahama-Tortugas region we find a Hydromedusan fauna composed of both Trachylina and Leptolina, the Leptolina forms of which are almost wholly distinct from those of the Canary Islands. We wish to call attention to the fact that a comparison of the Hydro- medusan fauna of the Bahama-Tortugas with that of the Canary Islands is open to serious objections, and that the conclusions arrived at through such a comparison may be of but little value. The Canary Islands occupy a small area, and are surrounded by water of 1000-2000 fathoms in depth, while the temperature of the surface water in their neighborhood is about 10° F. colder than that of the Bahama-Tortugas region. We might then expect that a marked difference would be observed in the Hydromedusan faunz of the two regions, for in the neighborhood of the Bahamas and Tortugas we find great areas of very shallow water having a very high temperature, while even the deepest parts of the Gulf of Mexico and Caribbean Sea have a temperature of 394° F. It would be much fairer and far more conclusive, were we able to do so, to institute a comparison between the fauna of the Bahama- Tortugas and that of the Gulf of Guinea in the neighborhood of the Islands of Anno Bom, St. Thomas, and Fernando Po; for here the con- tinental slope of the African coast is more gradual than at any other place, and the islands are surrounded by a depth of water not greater than 500 fathoms, having a bottom temperature of 393° F.; which is exactly the same as that of the deep parts of the Gulf of Mexico. The temperature of the surface water is also nearly the same as that of the Bahama-Tortugas region. The conditions at the Tortugas in August and September are very similar to those in the Gulf of Guinea 2 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. bo in February and March, as will become clear through an inspection of Kriimmel’s Temperature Charts (Kettler’s Zeitschrift, Bd. VI., Taf. IL, III. Also, Bull. Mus. Comp. Zodl. at Harvard Coll., Vol. XIV., pp. 240, 242, Figs. 168, 169). Unfortunately the Hydromedusan fauna of the Gulf of Guinea is unknown, but when we come to know it, we would not be surprised were it found that many Tortugas forms are established in this region. The Scyphomedus@ of the Bahama-Tortugas region are, for the most part, distinctly West Indian types, and are quite different from the species found on the Atlantic Coast of Africa. It is well known that these forms are much more abundant along coasts than they are in the open sea. The Discomeduse, especially, are given to congregating in swarms in bays and estuaries. We are therefore not surprised to find that most of the Bahama-Tortugas species are peculiar to the West Indies and the adjacent warm coasts of North and South America. Vanhoffen (1888 ; Bibliotheca Zoologica, and 1892; Acalephen der Plankton Expedition) has given maps showing the geographical dis- tribution of Scyphomeduse, and from an inspection of his charts it becomes quite apparent how these forms are distributed along coasts, and that few of them have yet been found in the open sea. Indeed, according to Vanhéffen (’92) only six Seyphomedusee were found by the Plankton Expedition of 1889, which confined its investigations, as far as possible, to the open sea far from coasts. The following Scyphomeduseze appear to be restricted to the Bahama- Tortugas region and the West Indies: Cassiopea frondosa, Lamarck ; Cassiopea xamacana, Bigelow; Linerges mercurius, Haeckel ; Linerges pegasus, Haeckel ; Linuche ungiculata, Eschscholtz; Linuche vesiculata, Haeckel ; Aurelia habanensis, Mayer ; Aurelia marginalis, L. Agassiz ; Charybdea xamacana, Conant; Tripedalia cystophora, Conant ; Charyb- dea punctata. In addition to these the following forms are established in the Bahama-Tortugas region, but extend also for a considerable distance northward along the coast of the United States: Pelagia cyanella, Péron and Lesueur ; Dactylometra quinquecirra, L. Agassiz ; Tamoya haplonema, F. Miiller. The following species extend from the West Indies southward along the Brazilian coast ; Dactylometra lactea, L. Agassiz ; Tamoya haplonema, F. Miiller. There are also a few Scyphomedusz of very wide distribution that are found in the region of the West Indies and Bahamas. Among these are: Nausithoé punctata, K6lliker, found in the Mediterranean, the Tropical Atlantic, and the Bahamas. A very close variety, N. MAYER: MEDUSZ FROM THE TORTUGAS, FLORIDA. 23 punctata var. pacifica, occurs in the Tropical Pacific. Periphylla hyacinthina, Steenstrup; found widely distributed throughout the whole Atlantic Ocean (see Vanhoffen, 1892; Akalephen der Plankton Expedition, Taf. V.). Pelagia phosphora, Haeckel ; appears to be widely distributed over the Tropical Zone of the Atlantic Ocean (see Haeckel, 1879, p. 507, Vanhoffen, 1892, pp. 19, 20); Atolla Bairdii, Fewkes, is a deep sea form that has been found by the “ Albatross” in the Gulf Stream, off the coast of the United States, and by Vanhoffen south of the Cape Verde Islands, off the African coast. The Siphonophore of the Bahama-Tortugas region are almost all widely distributed Tropical Atlantic forms, and most of them have already been found by Haeckel, and by Chun, in the Canary Islands. The Siphonophore are pelagic animals par excellence, and as they undergo their development while floating within the ocean, and are quite, if not wholly, independent of the bottom, one finds them widely distributed by ocean currents. As was pointed out by Chun (1897, Siphonophoren der Plankton Expedition, p. 101, etc.), the Siphonophorz of the warm regions of the Atlantic Ocean are widely distributed, distinctive species not being confined to particular regions. It is quite true, however, as Chun also shows (pp. 107-109), that, while many of the Atlantic Siphonophoree are found in the Mediterranean, there are others which are peculiar to the Mediterranean and have not been seen in the Atlantic; while there are also a number of Atlantic species that do not appear in the Mediter- ranean. It is possible, as future researches may demonstrate, that there are a few Siphonophore that are restricted to the Gulf of Mex- ico, or the Bahama Banks, but as yet we are certainly not justified in making any such statement. The Ctenophore of the Bahama-Tortugas region are not sufficiently well known, and too little has been discovered concerning their distribu- tion to warrant us in drawing general conclusions in regard to their geo- graphical range. Beroé Clarkii and Bolina vitrea appear to be confined to the West Indies and the southern Atlantic Coast of the United States, while Ocyroé crystallina probably has a wider distribution over the Tropical Atlantic. The so-called “ Eucharis multicornis,” “ Hormiphora plumosa,” and “ Beroé ovata” of the Tortugas have not been studied with sufficient care to warrant our stating that they are actually identi- cal with the Mediterranean species bearing the same names. 24 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. CoMPARISON OF THE Banama-TortuGAS FAUNA WITH THAT OF THE Fiz1 IshanpS AND TRopicaL PaciFic. In 1897, A. Agassiz and the author made a study of the Acalephian fauna of the Fiji Islands, South Pacific, in 18° 8. Lat., 178° E. Long. from Greenwich. The results of our investigations have been published in the Bulletin of the Museum of Comparative Zodlogy at Harvard Col- lege, 1899, and we there show that the Hydromedusz and Siphonophoree of the Fiji Islands are very closely related to those of the Tortugas, Florida. All of the Hydromedusz and Siphonophorz found by us in the Fiji Islands belong to well-known Atlantic genera. In the case of the Hydromedusze 4 Fijian species are so closely related to forms found at the Tortugas that we are unable to distinguish any specific differ- ence between them, and therefore we venture to assert that they may be identical species. These forms are Aiginella dissonema, Halitiara for- mosa, Pandea violacea, and AXquorea floridana. It will be observed that only one of these identical species belong to the Trachylina (¢. e. A. dissonema), the other three being Leptolina forms. In addition to the species already mentioned, the following genera of Hydromedusz are represented both in the Fiji Islands and in the Tortugas by very closely allied, although distinct species, — Aglaura, Eutima, Laodicea, Oceania, Epenthesis, and Tiaropsis. Among the Siphonophore (Abyla quincunx, Aglaisma quincunx) and Agalma Pourtalesii are found both at the Tortugas and Fiji Islands. Spheeronectes Kollikeri of the Fiji Islands and Tropical Pacific is cer- tainly very closely allied to Spheronectes gracilis of the Tortugas and Tropical Atlantic ; and the two species may eventually prove to be identical, and the same may be said of Nectophysa Wyvellei. The Scyphomeduse of the Fiji Islands are with two exceptions quite distinct from those of the Tortugas, for there are a number of | characteristic Rhizostomata in the South Pacific that have no near allies in the Atlantic Ocean. We find, however, in the Fiji Islands a variety of Nausithoé punctata that may prove to be specifically identi- cal with the form found at the Tortugas and in the Mediterranean. Another form, Linerges aquila, of Fiji is closely allied to, although distinct from, lL. mercurius of the West Indies. Among the Ctenophore of Fiji, Eucharis grandiformis is a species that bears quite a close resemblance to E. multicornis of the Atlantic and Mediterranean, although it is certainly specifically distinct. We must conclude, then, that the Acalephian fauna of the Fiji Islands is almost as closely related to that of the Tortugas as the latter is to * MAYER: MEDUSA) FROM THE TORTUGAS, FLORIDA. 25 that of the Canaries. It should be borne in mind, however, that the physical conditions in the Fiji Islands are in many respects quite similar to those of the Tortugas, and are very different from those of the Canary Islands. In both the Fiji and Tortugas Islands we find luxu- riant coral reefs and wide areas both of deep and shallow water, and in addition the temperature of the water in the two groups of islands is very nearly the same. In the Canaries, however, we find few corals, and no extensive shallow areas, the islands being surrounded by water of great depth. The temperature of the water there is also much lower than at the Fiji and Tortugas Islands. We have shown that the Tortugas medusz cannot survive in cold water, for not a single species is to be found upon the coast of New England north of Cape Cod. The Tortugas forms that are now established at the Fiji Islands must therefore have passed from the Atlantic into the Pacific Ocean somewhere within the tropical, or warm, regions of the Earth, and there can be but little doubt that the Tropical Atlantic was at one time in direct connection with the Pacific. Under these circumstances the Great Equatorial Current would pour from the Atlantic into the Pacific, and the pelagic life of the tropical regions of both oceans would become closely related. A fuller discussion of this subject, and of the researches of Hill, 1898 (Bull. Mus. Comp. Zo6l., Vol. 28) upon the geological history of the Isthmus of Panama will be found in our paper upon Fiji Acalephs in 1899. In view of the close relationship that exists between the Acalephian faunz of the Fiji and Tortugas Islands, one would be led to expect that the medusz of the Gulf of Panama and the west coast of Mexico would also display a resemblance to those of the West Indies and Tropical Atlantic; and this is, indeed, the case. Maas, 1897, in his report upon the meduse of the “ Albatross” expedition of 1891, records 18 species of Hydro- and Scypho- medusz belonging to 15 genera. All but one of the genera (Chiarella) are represented in the Atlantic by well-known species. Five of the Hydromeduse from the Gulf of Panama and Galapagos Islands are represented in the Atlantic by species so closely related to them that, were they found existing side by side in the same region, they would probably be considered to be varieties one of the other. Thus :— Stomotoca divisa, Maas ee { S. pterophylla, of the Bahamas. Homeonema typicum, Maas ee a HI. militare, of the Atlantic. Aglaura prismatica, Maas need related Liriope rosacea, Eschscholtz Geryonia hexaphylla, Brandt J L. cerasiformis, of the Atlantic. ! | { A. hemistoma, of the Atlantic. | | | G. (Carmerina) hastata, Mediterranean. 26 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. The following table will serve to show the wide geographical range of some species of Medusz found at the Tortugas, Florida. (0) indicates absence; (1) indicates that the species is identical with that found at the Tortugas. For example, (1) found in the column headed “ Canary Islands ” shows that the Canary species is identical with that found at the Tortugas. (1+) indicates the presence of a form that may prove to be identical with the Tortugas species. (A) indicates the presence of a closely allied but nevertheless distinct species from that found at the Tortugas. Tortugas, Canary Islands,| Fiji Islands, : Florida, Atlantic Ocean, |SouthPacific, | wediterranean Narici ch Species 24° 40’ N. Lat. | 28° 30’ N. Lat. | 18° S. Lat. Sea. 82° 53’ W. Long.| 15’ W. Long. |178° E. Long. Hydromeduse. /Eginella dissonema . 1 1 1 0 Aglaura hemistoma . 1 1 1: 1 Halitiara formosa . 1 0 1 0 Laodicea ulothrix 1 1 A 0 Pandea violacea 1 0 1 0 /Equorea floridana 1 0 EE 0 Staurodiscus tetrastaurus . 1 1 0 0 Tiaropsis heliosa . 1 0 A A Scyphomeduse. Nausithoé punctata 1 G 1+ 1 Siphonophore. Abyla pentagona . 1 1 0 1 Abyla quincunx il 1 1 0 Agalma Pourtalesii . 1 0 1 0 Diphyes bipartita. 1 1 0 1 Diphyopsis picta . 1 1 0 0 Physalia pelagica ; 1 1 0 1? Rhizophysa Eysenhardtii . 1 1 0 0 Rhizophysa Murrayana 1 1 0 0 Spheronectes gracilis 1 1 14 1 Ctenophore. Eucharis multicornis . . 12 u A 1 MAYER: MEDUS FROM THE TORTUGAS, FLORIDA. pay | Morpuoitoey or Tortucas MeEpus&. Among the new species described in this paper the following are worthy of special notice: Pseudoelytia pentata, « hydromedusa, is normally pentamerous, having 5 radial canals 72° apart, 5 gonads, and 5 lips to the proboscis. This curious species has probably been derived, philogenetically, from a pentamerous sport of some form of Epenthesis, and represents the survival of a discontinuous, meristic variation. Multioralis ovalis is a new genus of Hydromedusz in which 4 sep- arate manubria are situated upon a single straight chymiferous canal, which traverses the long diameter of the bell. Eucheilota paradoxa is the only Leptomedusa known which gives rise to young medusz by a direct process of budding. Niobia dendrotentacula isa remarkable form of Hydromedusa in which the tentacles develop into new meduse and are set free to propagate the species. This is accomplished through a process of growth, budding, and* fusion of parts. After all of the tentacles have been cast off, the adult medusa reproduces by a sexual process. In Bougainvillia niobe, Mayer, the medusa buds found upon the proboscis are formed entirely from the ectoderm, the entoderm taking absolutely no share in their construction. Oceania McCradyi of Brooks, 1888, a hydromedusa that produces hydroid-blastostyles upon its gonads, has been found at the Tortugas. In Dysmorphosa dubia, there appear to be 4 rudimentary gonads ? upon the 4 radial canals. If future observations confirm this conjecture, the case will be almost unique among Tubularian medusz. SumMMARY oF RESULTS. There is at the Tortugas, Florida, a tropical Medusan fauna, only three species of which are established upon the southern coast of New England ; and not one species of which is found upon the New England coast north of Cape Cod. The Hydromedusz of the Tortugas are more closely related to those of the Fiji Islands, South Pacific, than they are to those of the Canary islands, off the Atlantic Coast of Africa. In comparing the Hydromedusan fauna of the Tortugas with that of the Canaries, we see that the Leptolina forms of the Tortugas are almost wholly distinct from those of the Canary Islands. A number of Trachylina forms are, however, common to the two groups of islands. VOL. XXXVII.— No. 2. 2 28 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. As was shown by Maas, 1893, these Trachylina forms range widely over the open ocean ; and this observation has been confirmed by us during the cruise of the U.S. F. C. S. “Albatross” in the Tropical Pacific, 1899-1900. The Siphonophore of the Tortugas are very closely related to those of the Canary Islands. They also display a relationship to those of the Fiji Islands, South Pacific. The Scyphomedusz of the Tortugas are, for the most part, distinctly West Indian types, and are not closely related to forms known from the African coast. 33 Hydromeduse, 3 Siphonophore, 1 Hydroid, and 2 Scyphomeduse are new to science, and 44 forms are new to American waters. DESCRIPTIONS OF SPECIES. I HYDROMEDUS4. DIPURENA, McCrapy, 1857. Dipurena fragilis, nov. sp. Fig. 41, Plate 17. Specific Characters. — The bell is 4 mm. in height, and is half egg-shaped. The bell walls are of only moderate thickness. There are 4 long slender tentacles each bearing upon its distal end a single knob-shaped mass of nema- tocyst cells. A single black ocellus is situated in the ectoderm of the outer surface of each tentacle bulb. The velum is prominent. ‘There are 4 slender straight radial canals, and a narrow ring-canal. The proboscis is about 8 mm. in length, and exhibits two distinct annular swollen regions where the gonads are situated. The entoderm of the proboscis, and of the basal bulbs of the tentacles, is ochre-yellow. The entoderm of the distal nematocyst knobs of the tentacles is slightly orange. Several specimens were found at the Tortu- gas in June, 1897. This species differs from Dipurena strangulata of Charleston (see MeCrady, Proc. Elliott Soc., 1857, p. 38, Plate 9, Figures 1, 2) in that the tentacles are longer and much more slender ; and the color of the entoderm of the proboscis and tentacles is light ochre-yellow instead of rich green and red as in the Charleston species. ST MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 29 Dipurena picta, nov. sp. Figs. 45, 46, Plate 18. Specific Characters. — The bell is cylindrical in shape and 3 mm. in height. The bell walls are very thick and of a tough gelatinous consistency. There are 4 slender tentacles that are not quite as long as the bell height. These tentacles bear from 3-5 bulb-shaped nematocystic swellings near their distal ends (see Figure 46). The basal bulbs of the tentacles are large, and each one bears a dark purple ocellus. There are 4 straight radial tubes and a narrow circular tube. The velum is not very well developed. The proboscis is about 5 mm. in length and exhibits two distinct annular swellings that mark the places where the gonads are situated. The entoderm of the probos- cis and basal bulbs of the tentacles is of a beautiful custard-yellow. The entoderm of the nettle knobs of the tentacles is port-wine-colored. Two specimens were found at the Tortugas, Florida, during the first week in August, 1898. This species is closely allied to Dipurena dolichogaster, of the Mediterranean (see Haeckel, Syst. der Medusen, 1879, p. 25, Taf. II., Figures 1-7). It differs, however, from the Mediterranean form in that the bell is much thicker and more nearly cylindrical in shape, and there are fewer nematocyst-bearing bulbs upon the tentacles. STEHENSTRUPIA, Forszss, 1848. Steenstrupia gracilis, Brooks. Figs. 36, 37, Plate 16. Steenstrupia gracilis, Brooks, W. K., 1882, Studies Biol. Lab. Johns Hopkins Univ., Vol. II. p. 144. Specific Characters. —The bell is 4.5 mm. in height, and is surmounted by a slender apical projection fully 2 mm. in length. There are 2 rudimentary tentacle bulbs, one short, stiff tentacle, and one long tentacle which is ringed with a number of annular swellings. The velum is well developed. There are 4 slender radial canals and a narrow ring-canal. A long slender canal runs up from the proboscis into the apical projection of the bell. In mature speci- mens (Figure 36) the proboscis extends a short distance beyond the velar opening. The proboscis is cone-shaped, and the mouth is a simple round opening without oral lappets. The entoderm of the proboscis is intense yellow-green and rose-color. The entoderm of the tentacles is either yellow- green or rose-colored. Found at the Tortugas, Florida, and on the North Carolina coast, in July and August. 30 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. DINEMA, Van BeEnepey, P. J., 1867. Dinema jeffersoni,! nov. sp. Fig. 126, Plate 37. Specific Characters. — The bell is dome-shaped and higher than it is broad ; the height being about 1 mm. and the breadth 0.75 mm. The outer surface is sparsely sprinkled with nematocyst cells. There are 2 short marginal ten- tacles and 2 well-developed tentacle bulbs. The tentacles are covered with numerous small, wart-like, nematocyst-bearing swellings. The basal bulbs are well developed. There are 4 ocelli, one in each tentacle bulb. These ocelli are ectodermal and are situated on the centripetal sides of the bulbs. The velum is well developed. There are 4 straight narrow radial canals and a simple slender circular vessel. The proboscis is about as long as the height of the bell cavity. It is simple, round, and tubular, and the mouth-opening is situated at the extremity of a short cylindrical neck. A simple, short- style canal extends upward from the gastric cavity into the gelatinous sub- stance of the bell. The entoderm of the tentacles and tentacle bulbs is of a delicate green. The ocelli are bright red-brown, and the entoderm of the proboscis is flesh-colored. This form is occasionally met with at the Tortugas late in May and early in June. Dinema floridana, nov. sp. Specific Characters. —The bell is about 4 mm. in height and 3 mm. in diameter. The gelatinous substance is thin ‘and uniform, and the side walls of the bell are vertical. There are 2 well-developed, radially situated tenta- cles. Near the distal end of each of these tentacles there is a large knob- shaped swelling which terminates in a thin, nematocyst-bearing lash. The knob-shaped swelling is hollow and is connected with the general gastro-vas- cular system of the medusa by means of a narrow tube which extends through- out the length of the entodermal core of the tentacle. The basal bulbs are not large and there are-no ocelli. In addition to the 2 long tentacles there are 2 simple rudimentary tentacle bulbs 90° from the well-developed tentacles. The velum is well developed. There are 4 straight narrow radial canals. The proboscis is flask-shaped, being narrower at its base than at the middle of its length. It extends a short distance beyond the velar opening, and the mouth is a simple round opening, at the extremity of a long narrow neck. The entoderm of the proboscis and tentacle bulbs is bright yellow. The en- toderm of the swollen distal ends of the tentacles is yellow flecked with orange. A single specimen of this medusa was found at the Tortugas, Florida, June 17, 1897. 1 Named after Fort Jefferson, at the Tortugas, Florida. | . MAYER: MEDUSZ FROM THE TORTUGAS, FLORIDA. 31 HALITIARA, Fewses, 1882. Halitiara formosa, FEwKEs. Halitiara formosa, Fewkes, J. W., 1882, Bull. Mus. Comp. Zoél. at Harvard Coll., Vol. IX. p. 276, Pl. IV. Fig. 2. Specific Characters. — The bell is 3 mm. in height, and is provided with a solid apical projection. There are four long, radially situated tentacles, the distal ends of which are usually carried coiled in a tight helix. These tentacles are hollow, and have well-developed basal bulbs. In addition to these there are 24-35 short, solid tentacles that are usually carried tightly coiled. The velum is well developed. There are 4 straight, narrow radial tubes and a narrow, simple, circular vessel. The proboscis is pyriform, and extends for about half the distance from the apex of the bell cavity to the velar opening. The mouth is a simple round opening, and there are no prominent lips. The gonads are situated within the proboscis. In the case of the female the ova are very large and conspicuous. The entoderm of the proboscis and tentacle bulbs in the females is green; in the males, light brown. This medusa is cer- tainly the commonest of all at the Tortugas, Florida, during the summer months. We have found this species in the Fiji Islands. ; ECTOPLEURA, Aeassiz, L., 1862. Ectopleura minerva, nov. sp. Fig. 38, Plate 16; and Fig. 125, Plate 37. Ectopleura, sp, Fewkes, J. W., 1883, Bull. Mus. Comp. Zoél. at Harvard CollL, Vol. XL p. 85, Pl. I. Fig. 11. This form possesses but two marginal tentacles instead of four, as in all other species of Ectopleura. Specific Characters. — The bell is 2.5 mm. in height and is pear-shaped, hav- ing a well-developed apical projection. The gelatinous substance is of only moderate thickness. 8 rows of nematocyst cells extend from the tentacle bulbs to the bell apex. There are 2 well-developed tentacles, and 2 small tentacle bulbs. There are 6-9 separate, wart-like swellings upon the upper (aboral) side of each tentacle. These swellings are crowded with nettling cells. The velum is well developed. There are 4 straight, narrow, radial canals and a slender circular vessel. The proboscis is pear-shaped and is about 2 as long as the height of the bell cavity. A simple, short style-canal extends apie ard into the apical projection of the bell. The entoderm of the proboscis and tenta- cles is of a delicate purple, while the supporting lamella of the bell is of a on BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. decided green. There are a large number of brilliant yellow spots in the radial canals and tentacle bulbs. This form is rare at the Tortugas, Florida. It was found by Fewkes, 1883, at the Bermudas. STOMOTOCA, Agassiz, L., 1862. Stomotoca australis, nov. sp. Fig. 2, Plate 1. Specific Characters. —The bell is about 2.5 mm. in height, and there is a well-developed, solid, conical projection upon the aboral surface of the umbrella. The bell walls are thin. There are 2 large diametrically opposed tentacles, which are situated at the foot of two of the radial canals. The basal bulbs of these tentacles are hollow, and are long and conical. In addition to the two long tentacles there are two rudimentary tentacle bulbs situated at the bases of the radial canals 90° away from the long tentacles. There are 8 ectodermal ocelli. 4 of these are situated upon the centrifugal surfaces of the 4 tentacle bulbs, and the 4 others occupy intermediate positions upon the bell margin. The velum is wide. The radial canals and circular tube are broad, and their edges are smooth and simple. The proboscis is short and urn-shaped, and extends about halfway from the inner apex of the bell cavity to. the velar opening. The gonads occupy complexly folded and corrugated regions upon the sides of the stomach. The proboscis and tentacle bulbs are yellow, or greenish yellow. The ocelli are orange. The entodermal core of the proboscis often displays a faint orange tinge. This species is common throughout the summer at the Tortugas, Florida. Stomotoca rugosa. Stomotoca apicata, Fewkes, J. W., 1881, Bull. Mus. Comp. Zool., Vol. VIII. p. 152, PEE Hige 1, 4,0: Amphinema apicatum, Brooks, W. K., 1883, Stud. Johns Hopkins Biol. Lab., Vol. II. p. 473. This species has usually been confounded with Stomotoca apicata, L. Agas- siz. Stomotoca apicata, L. Agassiz, is, however, distinguished from S. rugosa by the circumstance that the entoderm of the proboscis in the male is green, or straw-colored, and in the female, dull ochre ; and the tentacle bulbs in the male are purple, and in the female, dull ochre. In the form described by Fewkes and Brooks, for which we propose the name S. rugosa, the entoderm of the proboscis and of the tentacle bulbs is always brick-red in both sexes, Specific Characters. —The bell is 5 mm. high and 3 mm. broad. It bears an apical projection which in some individuals is long and slender and in others short and blunt. The substance of this projection is solid throughout. well-developed tentacles and 14 small rudimentary ones. There are 2 long, MAYER: MEDUSZ FROM THE TORTUGAS, FLORIDA. 33 _ The basal bulbs of the long tentacles are large and hollow. When fully stretched, the long tentacles attain a length of 4-10 times the bell height. The velum is well developed. There are 4 broad radial tubes, and also a broad circular vessel with jagged outlines. The proboscis is flask-shaped, the lips being flanged and quite prominent. The mature sexual products are found in the ectoderm of the proximal portion of the proboscis where the outer surface is folded into a complex series of ridges. The bell is transparent. The ento- derm of the tentacle bulbs and of the proboscis is brick-red. In some indivi- duals the entoderm of the 4 radial tubes and of the circular vessel exhibits a faint tinge of red. The specimens of this species from the Tortugas, Florida, are peculiar in that the red color of the proboscis and tentacle bulbs is streaked with black. In some cases, after the medusz had been confined in aquaria for a number of days, the proboscis and tentacle bulbs became wholly black. This medusa is very common at Newport, Rhode Island, but does not extend north of Cape Cod. It is found all along the southern coast of the United States, but is rare at the Tortugas, Florida. Hydroid, and young medusa. — Brooks, 1883, describes the hydroid of this species. It is a Perigonimus, very much like P. minutus, Allman (1871 ; Tubularian Hydroids, p. 324, Plate XI. Figures 4-6). It was found growing upon the lower surface of the shell of Limulus, fastened to the sand tubes of Sabellaria. The stems are simple and unbranched and are about 0.2 mm. in height. The stems are covered for about two thirds of their length by a deli- cate, closely adherent film of perisare to which foreign particles become attached. The stomach occupies about one fourth or one fifth of the stem, from which it is separated by a slight constriction. Each polypite possesses ten tentacles, which point alternately backwards and forwards, those pointing forwards being a little longer than the others. The medus@ are attached by very short peduncles to the sides of the stems. When the medusa is set free it is about 0.5 mm. in height, and there is no trace of the apical projection, which develops in the course of about 8 days. In an abnormal individual of this species found at Newport, Rhode Island, in July, 1892, there were 4 long tentacles, one at the base of each radial canal. This medusa was maintained alive in an aquarium for more than a month. When first found it had but two diametrically opposed tentacles, each at the foot of a radial canal. The other pair of large tentacles developed later, after the first pair had attained their full length. The medusa then possessed 4 long tentacles and 12 small rudimentary tentacle bulbs. This variation is interest- ing, as it illustrates the close relationship between Stomotoca and Modeeria. 34 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. PANDBEA, Lesson, 1837. Pandea violacea, AGassiz and Mayer. Fig. 1, Plate 1. Pandea violacea, Agassiz, A., and Mayer, A. G., 1899, Bull. Mus. Comp. Zodl. Harvard Coll., Vol. XXXII. p. 160. Specific Characters. — The bell is pear-shaped and 4 mm. in height. The bell walls are only of moderate thickness. There are 32 tentacles, 8 of these are each about 3 times as long as the bell height, and 24 are small and rudi- mentary. The basal bulbs of the long tentacles are hollow. There are 32 ocelli, one on each tentacle bulb. The velum is well developed. The proboscis is flask-shaped, its proximal portion being distended by the 4 gonads. The lips are simple and cruciform. There are 4 straight radial tubes, and a broad circular vessel. The entoderm of the proboscis and tentacle bulbs is of a deli- cate pink. A green streak runs along the outer surface of the entodermal lining of the radial canals. The ocelli are purple in color. The medusa is common at the Tortugas, Florida, throughout the summer. We have found a species at the Fiji Islands that appears to be identical with the Tortugas form. Our figure is drawn from a specimen found at the Tortugas. TIARA, Lesson, 1843. Tiara superba, nov. sp. Fig. 39, Plate 16. Specific Characters. — The bell is 5 mm. in height and possesses a small apical projection. There are 4 long hollow tentacles and 12 small rudimentary tentacles. A brilliant red eye-spot is found in the ectoderm of the outer sur- face of each tentacle bulb. The velum is well developed. There are 4 broad straight-edged radial tubes and a broad circular vessel. The proboscis is very broad and the lips are surrounded by complexly fimbricated lappets. The gonads are found in 4 sharply folded, radially arranged regions in the upper portion of the proboscis. The proboscis is bound to the radial tubes by means of 4 mesenteries. The entire gelatinous substance of the medusa is of a deli- cate rose-pink. The entoderm of the proboscis and tentacles is of a rich rose- color, and the entodermal core of the proboscis is emerald-green. This medusa makes its appearance in June and continues to be common throughout the summer at the Tortugas, Florida. MAYER: MEDUSA! FROM THE TORTUGAS, FLORIDA. vi GEMMARIA, McCrapy, 1857. Gemmaria dichotoma, nov. sp. Fig. 40, Plate 17. Specific Characters. — The bell is 3 mm. in height and there is a solid mitre- shaped apical projection. The bell walls arethin. There are two rudimentary tentacle bulbs and two well-developed tentacles. The entodermal core of these large tentacles is hollow. They terminate in a bulb-shaped nematocyst swelling, which in some individuals is provided with delicate bristles. A number of tentacule arise from the upper or ‘‘ dorsal” side of the tentacle, and each one of these terminates in a bulb-shaped swelling similar to that at the distal end of the main tentacle. The youngest and least-developed of these side branches is always found nearest the bell. The basal bulbs of the tentacles are large, and there is a single deep red ocellus in the outer surface of the ectoderm of each. The velum is quite well developed. There are 4 straight radial canals and a narrow circular canal. The proboscis is pyriform and extends about half the distance from the apex of the bell cavity to the velar opening. The entoderm of the proboscis and tentacles is ochre-yellow. Several specimens were found at the Tortugas early in July. Gemmaria gemmosa, McCrapy. Figs. 137, 138, Plate 41. Gemmaria gemmosa, McCrady, J., 1857, Gymn. Charleston Harbor, p. 49. Zanclea gemmosa, McCrady, J., 1857, Gymn. Charleston Harbor, p. 48, Pl. 8, Figs. 4, 5. Specific Characters. —Hydroid stock ; Gemmaria gemmosa. The hydroid was found at the Tortugas, Florida, growing upon a piece of floating gulf- weed (Sargassum). The hydrorhiza is creeping and net-like, and gives rise at irregular intervals to short, more or less twisted hydrocauli. Both the hydro- rhiza and hydrocauli are covered with a horny, chitinous perisare, which in the hydrocaulus displays a number of annulations. The hydrocaulus is corru- gated, and opaque in color, throughout its length; and in this respect differs from the European G. implexa described by Allman (1871, Tubularian Hydroids, p. 290, Plate VII.). The fully developed hydranths are only 1.5 mm. in height. They are elongate, and the diameter near the proximal end is a little greater than at the free oral extremity. The tentacles arise in 5-8 whorls from the side of the hydranth. Each whorl contains 4-6 short tentacles. Each tentacle terminates in a distal knob which is armed with a dense cluster of nematocysts. The cells of the shafts of the tentacles are vacuolated, and the tentacles themselves quite stiff and inflexible. 4-8 medusa-buds arise from the side of the hydranth immediately below the proximal whorl of tentacles. 36 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. When set free the young medusa possesses 2 well-developed diametrically opposed tentacles and 2 rudimentary tentacle bulbs (Figure 137). The 4 radial, nematocyst-bearing swellings upon the ex-umbrella extend halfway up the sides of the bell from the margin toward the apex. The bell walls are uniform, and very thin and flexible. There are 4 slender radial canals, and the proboscis is a short simple tube with no trace of gonads. Before being set tree, the tentacles are carried coiled inward so that they lie protected within the bell cavity. Soon after liberation, however, the tentacles are turned out- ward. (Compare Figures 137 and 138.) The deep-lying entoderm of the hydranth is of a delicate creamy pink, while the more superficial entoderm is of a translucent milky color. The entodermal cells of the superficial entoderm are large and vacuolated. The hydrorhiza is of a horny yellow color. This species is quite different from Gemmaria implexa of Allman. It is probably the hydroid of Zanclea gemmosa, McCrady, of Charleston Harbor, but not having been able to raise the medusee we must remain in some doubt concern- ing its identity. NIOBIA, nov. gen. Niobia dendrotentacula, nov. sp. Figs. 141-143, Plate 42; Fig. 144, Plate 43. Generic Characters. — Niobia. Cladonemide with 2 simple and 2 bifurcated radial canals. There are 4 simple lips to the proboscis, but no oral tentacles. The marginal tentacles develop into free-swimming meduse. There is no place in the system of Haeckel (1879; p. 101) for this genus. It cannot be placed among the Dendronemide, for it has no oral tentacles, and as it has branched radial canals it cannot be classed among the Pteronemide. It combines the essential characters of both of these subfamilies, however, and forms a good connecting link between them. Specific Characters. — Adult medusa. The bell is slightly flatter than a hemisphere, and is about 4 mm. in diameter. The gelatinous substance is quite thin and uniform, but not very flexible. The tentacles are arranged in bilateral symmetry, the axis being in the diameter of the two simple radial canals (see Figure 144, Plate 43). The oldest tentacle is situated at one end, and the youngest at the other end of this axis. (Figures 142,144.) Each half of the medusa is a reflection of the other, and the order in age of the tentacles is given by the following diagram, the oldest tentacle being numbered (1) and the youngest (7) :— MAYER: MEDUSA! FROM THE TORTUGAS, FLORIDA. 37 7 4 a 2 2 6 6 3 3 5 5 1 Tentacles (1) and (7) are situated at the bases of the simple radial canals, while tentacles (2, 2) and (3, 3) are found at the bases of the two bifurcate canals. In addition to these there are the intermediate sets of tentacles (4, 4), (5, 5), and (6, 6); and thus the medusa possesses 12 tentacles, each succes- sive pair being 30° apart. It is very remarkable that through a peculiar pro- cess of growth each tentacle bulb is developed into a young medusa which resembles the adult, and is finally set free into the water. Various stages of this process will be seen by an inspection of Figures 141, 142, and 144. The oldest tentacle is the first to be transformed into a new medusa, and the others follow in the order of their age until all of the tentacles have been cast off. The first stage in this process is the development of a hernia-like outgrowth, involving both entoderm and ectoderm, adjacent to and on the centripetal side of each tentacle bulb upon the floor of the sub-umbrella. Soon after this two pointed outgrowths appear on both sides of each tentacle bulb, and finally develop into new tentacles. These outgrowing tentacles become larger, and soon a still younger pair make their appearance centrifugal to the first, and these are soon followed by two others which lie centripetally from the oldest pair. Before this, however, 4 short canals (the radial canals of the future medusa) develop, and place the gastric cavity of the future proboscis into com- munication with the circular vessel. An opening then appears in the velum of the adult medusa immediately below the proboscis of the future medusa, and this constitutes the velar opening of the new animal. The proboscis becomes cruciform in cross-section, and finally the new medusa is constricted off and becomes free in the condition represented in Figure 142. Here we see that the simple radial canals, the circular canal, the velum, and the oldest tentacle are stolen directly, so to speak, from the parent medusa. The forked canals, proboscis, and younger tentacles are new growths. Even before the outgrowing medusa is detached from the old one, hernia-like outgrowths appear upon the 38 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. sub-umbrella wall near the bases of its tentacles, and thus the process of form- ing new medusz is repeated in the next generation. The meduse are very hardy when detached and grow rapidly, and proceed at once to develop new meduse from their own tentacle bulbs. When detached, the bell of the new medusa is about 1.5 mm. in diameter. It is difficult to comprehend the phi- logenetic history of this curious and fortuitous combination of local growth, fusion, and budding which results finally in the formation of a medusa exactly resembling the adult. It is probable, however, that it has been derived from the usual budding process so common in hydromeduse, but that in this case a greater and greater number of parts have been taken directly from the adult medusa, until the present state has been arrived at. After the original tenta- cles have been cast off, new ones grow out in their places, and thus the old medusa always has 12 tentacles. After every one of the original 12 tentacles has been cast off, however, the process of forming new meduse becomes less active and finally ceases altogether. Then the gonads develop in 4 separate interradial regions on the wall of the gastric part of the proboscis. In the female the ova become very prominent, and are finally dehisced into the water. I was unable to raise them, however, and know nothing of the development of the sexual generation. The proboscis is flask-shaped, and there are 4 simple cruciform lips. The entoderm of the proboscis tentacle bulbs and circular canal is ochre-yellow, all other parts of the medusa being transparent. The medusa is very active and thrives well in confinement. Large numbers of them appeared at the Tortugas, Florida, on May 21, and continued more or less common until June 4, 1899. TURRITOPSIS, McCrapy, 1857. Turritopsis nutricula, McCrapy. Turritopsis nutricula, McCrady, J., 1857, Gymn. Charleston Harbor, p. 25, Pls. VeVi. MLL nip Modeeria multitentacula, Fewkes, J. W., 1881, Bull. Mus. Comp. Zoél. Harvard Coll., Vol. VIII. p. 149, Pl. ITI. Figs. 7-9. This medusa was well described by McCrady in 1856 and 1857. Fewkes, 1881, however, redescribed it as a new species under the name ‘ Modeeria multitentacula.” To add to the confusion respecting this species, a medusa that has since been identified by Martha Bunting, 1894, as Podocoryne carnea, was described by A. Agassiz, 1862, 1865, under the name of “Turritopsis nutricula.” The latter author was deceived by the close resemblance of the young of Podocoryne carnea to the young medusa of T. nutricula, MeCrady, 1857, into the belief that the two were identical. The mature meduse, how- ever, are easily distinguished one from the other, and the hydroid stocks differ widely from each other. TS pte Sl ar MAYER: MEDUS# FROM THE TORTUGAS, FLORIDA. 39 Specific Characters. — Mature medusa. The bell is pear-shaped with thin walls, and is 4mm. in height. There are 40-50 marginal tentacles that are capable of much contraction and extension. There is a single brown, ecto- dermal, pigment spot upon the centripetal side of each tentacle near the point of its junction with the tentacle bulb. The velum is well developed. There are 4 straight, narrow, radial canals. The proboscis is wide and fills about half of the cavity of the bell. The upper portion of the proboscis consists of highly vacuolated cells, or chambers, through the midst of which run the 4 radial canals. The mouth opening of the proboscis is found at the end of a short, narrow, cylindrical neck, and is surrounded by 4 radially arranged nematocyst-bearing knobs. The gonads are situated within the proboscis. The entoderm of the proboscis is dull yellow, streaked with brownish orange. The ocelli of the tentacle bulbs are orange, or brown in color. This medusa is extremely abundant from the coast of Cuba to Newport, Rhode Island. It is not found north of Cape Cod, Massachusetts. It is very common in Charleston Harbor, South Carolina, where it is infested by the young of Cunoctantha octonaria. This medusa is one of the few that ap- pears to develop from the hydroid stock both at the Tortugas and at New- port, Rhode Island. For while meduse indigenous to the Tortugas are often driven into Newport Harbor by southerly winds, very few of these southern visitors establish themselves permanently in the northern waters. The hydroid stock of this species was found by Brooks, 1886, at Morehead City, North Carolina. It is a Tubularian belonging to the genus Dendroclava. Brooks gives a number of good figures of it in his paper in the Memoirs of the Boston Society of Natural History, Vol. III., 1886. CYTAHIS, Escuscuorrz, 1829. Cytaeis gracilis, nov. sp. Figs. 122-124, Plate 36. Specific Characters.— Mature medusa; Figure 122. The bell is dome- shaped and a little broader than it is high, and the aboral apex terminates in a slight projection. The animalis3mm. in diameter. The gelatinous substance of the bell is of only moderate thickness. There are 8 quite stiff curled tenta- cles; 4radialand 4interradial. The radial tentacles are about two thirds as long as the bell height, while the interradial ones attain only about one half this length. The basal bulbs of all of the tentacles are large and deeply pigmented. The velum is broad. There are 4 straight, narrow, radial canals, and a simple slender circular vessel. The proboscis is mounted upon a short, wide peduncle. The gastric portion of the proboscis is only about one half as long as the height of the bell cavity. The mouth is a simple, round opening, surrounded by 8 un- branched oral tentacles. 4 of these tentacles are radial and 4 interradial in position, and each one terminates in a knob-like end formed of spindle-shaped 40 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. nematocyst cells. Medusa buds arise from the sides of the gastric portion of the proboscis. When set free each medusa possesses 4 short equally developed tentacles. The entoderm of the proboscis is usually red in color, although sometimes the inner core is red or pink and the outer annulus is green. Each tentacle bulb is provided with a dense mass of red entodermal pigment, which in the case of the radial tentacles extends for a considerable distance up the course of the radial canals. Young Meduse. — The youngest medusa found free in the water was 1.5 mm. in height (Figure 124), and the bell was about twice as high as it was broad. The gelatinous substance was quite thin. There were only 4 mar- ginal tentacles and these were radial in position. The distal tips of these ten- tacles were slightly knobbed and their entoderm was tinged with green. There were 8 simple oral tentacles, and the proboscis lacked a peduncle. In an older individual, which was 2.5 mm. in height, the bell was pyriform. The proboscis possessed a peduncle, and there were 8 marginal tentacles, 4 radial and 4 interradial. There were no traces of medusa buds upon the proboscis. This medusa was-quite common at the Tortugas, Florida, about the middle of June, 1899. DYSMORPHOSA, Puuieri, 1842. Dysmorphosa dubia, nov. sp. Figs. 64-66, Plate 22. Specific Characters. — The bell is egg-shaped and 1.5 mm. in height. The bell walls are thin and flexible. There are 8 quite stiff tentacles (Figure 66) that are carried curled slightly upward. The distal ends of these tentacles are thickly covered with nettling cells. A very large black ocellus is situated in the ectoderm of the under side of each tentacle bulb. The velum is well developed. There are 4 straight, narrow, radial tubes ; and a slender circular canal. The proboscis is pear-shaped, and there is a slightly developed peduncle. 4 radially situated oral tentacule surround the mouth. Each one of these terminates in a knob-like cluster of nematocysts (Figure 65). 4 small, rudi- mentary gonads? appear to be developed at points midway along the lengths of the 4 radial canals. The entoderm of the proboscis, tentacle bulbs, and gonads ? is of a delicate yellow. One specimen was found at the Tortugas, Florida, on July 20, 1898. The presence of what appear to be gonads ? upon the radial tubes is certainly remarkable ; it should be remembered, however, that such appearances are not unknown among genera of Tubularian meduse that normally bear their gonads upon the proboscis. In the case of Dipurena halterata bodies that are very similar in general appearance to rudimentary gonads are found upon the radial canals. (See Forbes, E., 1848, British Naked-Eyed Medusa, p. 53, Plate VI., Figures 1, b, c, d. Also Browne, E. T., 1898, Proc. Zool. Soe. London, p. 816, Plate 49, Figure 2. ) MAYER: MEDUSA) FROM THE TORTUGAS, FLORIDA. 41 Dysmorphosa minuta, nov. sp. Fig. 42, Plate 18. Specific Characters. —The medusa is extremely minute, the bell being only 0.3 mm. in height. It is pear-shaped and the walls are quite thick. The ge- latinous substance is remarkably delicate, and the medusa soon contracts into a shapeless mass in captivity. There are 8 marginal tentacles, with well- developed basal bulbs. The velum is small. There are 4 straight, slender radial canals and a narrow circular vessel. The proboscis possesses a distinct peduncle. The gastric portion as well as the peduncle is 4-sided in cross- section. 4 well-developed oral tentacles surround the mouth, one being situated at each radial corner. Each of these tentacles terminates in a knob-shaped distal end, which is thickly covered with nematocysts. The entodermal cells of the oral tentacles are disk-shaped and highly vacuolated. Several medusa buds in various stages of development are found upon the upper interradial regions of the gastric portion of the proboscis. In some specimens the ento- derm of the proboscis and tentacle bulbs is turquoise blue, and in others lilac. The medusa was common at the Tortugas, Florida, in the middle of July, 1898. It is the smallest hydromedusa known. Its color is also very different from D. fulgurans, A. Agassiz, of Newport Harbor. BOUGAINVILLIA, Lesson, 1836. Bougainvillia frondosa, nov. sp. Fig. 5, Plate 3. Specific Characters. —The bell is dome-shaped and about 2 mm. in height. There are 4 bunches of marginal tentacles, which are situated at the bases of the 4 radial canals. Each tentacle bulb gives rise to but 2 tentacles, thus making 8 in all. There are no ocelli at the bases of the tentacles. The velum is small, There are 4 straight, simple, radial tubes. The proboscis is short, thick, and flask-shaped, and extends only about one half of the distance from the inner apex of the bell cavity to the velar opening. There are 4 radially situated oral tentacles, each of which branches dichotomously two or three times. The mature gonads are found in 4 radially situated swollen regions upon the ectoderm of the proboscis, above the origins of the oral tentacles. ‘There are a number of flask-shaped bodies with narrow necks protruding from the surface of the proboscis in the region of the gonads. Each of these flask- shaped capsules is filled with yellow-colored cells. Although it is possible that these may be developing planule, we incline to the opinion that they are parasitic zo6xanthellz. We are led to this opinion on account of the decided yellow-green color of these cells, and also because we have found similar cap- sules scattered irregularly over the surface of the sub-umbrella of Laodicea 42 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. ulothrix at the Tortugas. The entoderm of the proboscis and tentacle bulbs is cream-colored, and the tips of the tentacles are turquoise. A single specimen of this medusa was found at the Tortugas, Florida, on June 11, 1897, and another in June, 1899. Bougainvillia niobe, Mayer. Bougainvillia niobe, Mayer, A. G., 1894, Bull. Mus. Comp. Zoél. at Harvard Coll., Vol. XXV. p. 286, Pl. I. Fig. 2. Specific Characters. — The bell is 6.75 mm. in height and 4.8 mm. broad. The bell walls are thick and gelatinous. The marginal tentacles arise from 4 radially situated bulbous swellings, each one of which gives rise to 6-8 ten- tacles. At the base of each tentacle, upon the inner or centripetal side, there is a dark-colored pigment spot, or ocellus. This is an ectodermal structure, and it projects slightly from the surface of the tentacle. The tentacles are not very flexible and are about as long as the bell height. The velum is well developed. There are 4 straight, narrow, radial tubes. The proboscis is wide, but not very long, extending only about half the distance from the apex of the bell cavity to the velar opening. There isa small peduncle. There are 4 large, radially arranged bunches of oral tentacles. These arise as 4 main stems, each of which branches dichotomously 4 times, thus giving rise to 16 tentacle tips from each quadrant of the proboscis. These terminal tentacle tips are slightly knobbed, and are composed chiefly of nematocyst cells. The tentacles of the proboscis are very flexible and may be observed waving gracefully to and fro within the cavity of the bell. The most remarkable characteristic of this species is the presence of numerous medusa buds that arise from the gastric region of the proboscis. These budding meduse are found in 8 radially arranged clusters situated near to and on both sides of the places where the 4 radial tubes enter the gastric portion of the proboscis. A study of sections of the proboscis of meduse killed in Flemming’s Chrome-Osmic-Acetic, and stained in Kleinenberg’s 70% Alcoholic Hematoxylin, has shown that the proliferating meduse are formed entirely from the ectoderm, the entoderm taking no part whatsoever in their formation. There is a very well-defined lamella between the ectoderm and the entoderm of the proboscis of the parent medusa, and the membrane of this lamella is never broken during the time of the formation of the medusa bud from the ectoderm of the proboscis. Indeed, the gastro-vascular cavity of the budding medusa is never connected with that of the parent. The medusa buds develop very much as has been demonstrated by Chun (1895 ; Bibliotheca Zoologica, Heft 19, Lfg. 1, p. 1-51, Taf. I., II.) in Rathkea octopunctata, and Lizzia Claperédei ; excepting that while in the forms studied by Chun the gastro-vascular cavity of the bud finally acquires a connection with that of the parent, in Bougainvillia niobe no such connection is ever formed. Chun concluded that medusa buds which are derived entirely from ectoderm cannot be homologous with those that are MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 43 formed from both ectoderm and entoderm in the manner commonly observed in Hydroids, and in the medusa of Sarsia ; for it is necessary, if organs be homologous, that they have a similar origin. It has occurred to us, however, that Chun may be mistaken in this conclusion, and that his statement may be more a matter of definition than of fact; for it may well be that, in the course of phylogeny, the entoderm has come to take less and less part in the formation of medusa buds, until finally, as in the case of Bougainvillia niobe, it has abandoned all share in their formation. Considered from the physiological standpoint it may be that in B. niobe the ectoderm of the parent proboscis being very thick, there is an abundance of cells from which to form the bud without having resource to those of the deep-lying and somewhat inaccessible entoderm. When set free the young medusa possesses 4 radial tentacles. The bell of the medusa is transparent, and the entoderm of the proboscis and ten- tacle bulbs is rosin-yellow. Found in Nassau Harbor, New Providence Island, Bahamas, in March, 1893. It is interesting to notice that Hartlaub (1897; Hydromedusen Helgolands) has shown that the sex cells of Bougainvillia superciliaris are first found in the entoderm of the young medusa, and that as development proceeds they pass into the ectoderm, where they become mature. It is possible that the cells which give rise to the medusa buds of Bougainvillia niobe are similarly derived from the entcderm of the young medusa. We have not seen the young and ~ immature medusa of B. niobe, and in the mature animal the supporting la- mella between the ectoderm and entederm of the proboscis is very distinct and unbroken, and we have never succeeded in discovering any cells which were passing through it. Margelis carolinensis, Acassiz, L. Hippocrene carolinensis, McCrady, J., 1857, Gymn. Charleston Harbor, p. 62, Pl. 10, Figs. 8-10. Margelis carolinensis, Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV. p. 344. In the Tortugas and Charleston Harbor examples of this species, the ento- derm of the tentacle bulbs and of the proboscis is of a delicate sage-green color ; while the gonads are cream-colored, and the tentacular ocelli dark-brown or black. Innorthern examples of this medusa, found at Newport, Rhode Island, and Naushon, Massachusetts, A. Agassiz describes the color of the tentacle bulbs as brilliant red surrounded by a green edge bordered with light yellow; and the digestive cavity as brick-red, or green. No such brilliant coloration has been seen in the southern specimens. The medusa is not very common at the Tortugas, being met with only occasionally during the summer months. In Charleston Harbor, South Carolina, however, it is extremely abundant. VOL. XXXVII. — NO. 2. 4 44 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. LIZZIA, Forsss, 1846. Lizzia elegans, nov. sp. Fig. 127, Plate 38, Specific Characters. — The bell is 3-7 mm. in height, and the gelatinous substance is of moderate and nearly uniform thickness. The sides of the bell are almost straight and vertical, and the top is dome-shaped. There are eight groups of marginal tentacles, 4 radial and 4 interradial. ach radial group is composed of 4, and each interradial of 3 tentacles. The tentacles are quite stiff and curve upward, and are only about one-half as long as the bell height. There is a small dark-brown ectodermal ocellus upon the under (oral) side of each tentacle near the basal bulb, The velum is wide and provided with strong muscles. There are 4 straight, narrow, radial canals, and a simple circular vessel. There is a well-developed conical peduncle to the proboscis, down which the radial canals lead in their course to the gastric sac. The gastric part of the proboscis is cruciform in cross-section and pear-shaped in general longitudinal contour. The mouth is a simple round opening without prominent lips. The oral tentacles arise from the four radial sides of the proboscis at a short distance above the mouth. ach tentacle branches dichoto- mously 3 times and then each tip terminates in three small branches which are covered with nematocysts. The entodermal cells of the oral tentacles are chordate. The genital products are situated upon the 4 radial sides of the gastric portion of the proboscis. The entoderm of the tentacle bulbs and radial canals is of an intense opaque pearly-white color, often displaying a tinge of pink. The entoderm of the gastric portion of the proboscis is of an intense green, and the oral tentacles are pearly-pink. The intense opaque color of the radial canals contrasting with the hyaline transparency of the bell renders this medusa one of the most beautiful to be found at the Tortugas, Florida. Several specimens were captured early in July, 1899. DISSONEMA, Haecket, 1879. Dissonema turrida, nov. sp. Figs. 3, 4, Plate 2. Specific Characters. — Adult medusa ; Figure 3. The bell is about 4mm. in height. Itis blunt and cone-shaped, and there is a prominent apical projection, which is hollow. There are 2 large hollow tentacles, which when expanded are 3-4 times as long as the bell height. In addition to these, there are 14 small solid tentacles, or marginal cirri. There are 16 ocelli, one at the base of each tentacle. These ocelli are situated within the ectoderm of the outer MAYER: MEDUSA! FROM THE TORTUGAS, FLORIDA. 45 (centrifugal) side of the tentacles. The proboscis is pyriform, and the lips project beyond the velar opening. The walls of the proboscis are very thin, and the lips are crenulated. The 4 radial canals are broad, and the 4 gonads occupy their proximal halves. In the female each gonad contains about six large ova, which stand out prominently over the surface of the organ. The entoderm of the proboscis and tentacles is of a delicate shade of green. The genital organs and circular canal are tinged with pink. Young Medusa. — Figure 4, Plate 2, represents a young medusa of this species in which the genital organs have not yet made their appearance. There are but 4 tentacles, and 8 ocelli; and it is remarkable that the long tentacles are as yet solid, although they become hollow throughout their length in the adult medusa. This species is common throughout the summer at the Tortugas, Florida. NETOCERTOIDHES, nov. gen. Netocertoides brachiatum, nov. sp. Figs. 43, 44, Plate 18. Generic Characters. — Cannotide with 8 bifurcating, radial canals. 16 canals reach the circular vessel. There are neither marginal sense-organs nor cirri. Specific Characters. — The bell is mitre-shaped and 3 mm. in height. There are 32 marginal tentacles. 16 of these are well developed, and are situated at the bases of the 16 radial canals; and the others are smaller, and alternate with the large tentacles in position. The large tentacles are only about one quarter as long as the bell height, and the others are much smaller. There are no marginal sense-organs. The velum is well developed. The proboscis has the shape of an 8-rayed star, each ray of which bifurcates, thus giving rise to 16 radial canals which reach the circular vessel. The gastric portion of the proboscis is wide, but flat, and the mouth extends but a short distance down into the bell cavity. The gonads appear to be situated upon the 8 rays of the stomach. Two specimens were found at the Tortugas, Florida, on July 10, 1898. As it floats in the water this medusa bears a wonderfully close resemblance to the little pelagic Alga (Trichodesmium), which is very abundant at the Tortugas. 46 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. STAURODISCUS, Harcket, 1879. Staurodiscus tetrastaurus, HAEcKEL. Figs. 47-49, Plates 18, 19. Staurodiscus tetrastaurus, Haeckel, E., 1879, Syst. der Medusen, p. 145, Taf. IX. Figs. 1-3. Specific Characters. —The bell is 4.5 mm. in diameter, and about twice as wide as it is high. In adult medusz there are 8 long flexible tentacles with hollow basal bulbs. In some specimens there are 24, and in others 16 sensory clubs upon the bell margin. There are always 32 black entodermal ocelli, one at the base of each tentacle bulb and sensory club. The velum is well developed. Only 4 radial canals reach the circular vessel. Each of these canals gives rise to a pair of side branches that end blindly. The gonads are situated upon these side branches and upon the distal portion of each radial canal. The proboscis bears 4 prominent lips. The color of the entoderm of this medusa is green or yellow. In the youngest specimen observed, the bell was 1 mm. in diameter and about as high as it was broad. There were 4 well-developed tentacles, 4 rudi- mentary tentacle bulbs, and 8 marginal clubs (see Figure 47, Plate 18). The medusa was very common at the Tortugas, Florida, in July and August, 1898. Haeckel, 1879, found this species in the Canary Islands, at Lanzerote. TETRACANNOTA, nov. gen. Tetracannota collapsa, nov. sp. Figs. 14-16, Plates 7, 8. Generic Characters. — Tetracannota is closely allied to Cannota and Bere- nice. It may be defined as having 16 radial canals, which in the adult become arranged in 4 groups, each group consisting of 4 canals. Gonads 16 in num- ber, and situated upon the distal regions of the radial canals, An entodermal pigment spot at the base of each tentacle. No otocysts. Tentacles numerous. Specific Characters. — Adult medusa; Figure 14. The bell is 7 mm. in diameter, and about as high as it is broad. The top is dome-shaped, and the side walls are vertical. There are 16 well-developed tentacles that are carried tightly coiled in close helices. In addition to these there are 112 very small, rudimentary tentacles. Dark-brown entodermal pigment is found at the base of each tentacle. There are 16 radial canals, arranged in 4 groups of 4 each. The gonads are found in the proximal portions of the 16 radial canals very near to the point where they branch off from the proboscis. The peduncle of the proboscis is wide and prominent. The proboscis possesses 8 slightly erenu- i -_— MAYER: MEDUSZ FROM THE TORTUGAS, FLORIDA. AT lated lips. The entoderm of the proboscis in some specimens is green, in others pearly-white or yellowish. The entodermal pigment spots at the bases of the tentacles are dark brown. Stages in Development.— The youngest medusa observed possessed a bell 1.5 mm. in diameter (see Figure 15). It had 4 simple radial canals, and 32 tentacles, 4 well developed and 28 rudimentary. The velum was prominent. There were 4 lips to the proboscis, and as yet no peduncle. There was no trace of the genital organs. In the next older stage (Figure 16), we find 16 radial canals, and 8 lips to the proboscis. As yet there is no peduncle and no trace of the gonads, nor have the radial tubes grouped themselves into four bundles as in the adult. This medusa was very common at the Tortugas in June, and ample oppor- tunity for observing its transformation was afforded. It possesses the curious habit of collapsing into an almost shapeless mass, in which condition it may remain for several hours and then “straighten out” and swim about in ex- cellent condition. Fewkes, 1883 (‘On a Few Meduse from the Bermudas,” Bull. Mus. Comp. Zodl., Vol. XI., No. 3, Figures 7, 79) has evidently figured the young of this species under the name of “ Larva of an unknown Tubularian.” DYSCANNOTA, Haecket, 1879. Dyscannota gemmifera. Fig. 17, Plate 8. Willia ornata? Brooks, W. K., 1880, American Naturalist, Vol. XIV. p. 670. Willia ornata, Brooks, W. K., 1881, Studies Johns Hopkins Univ. Marine Lab., Vol. II. p. 144. Wiilia gemmifera, Fewkes, J. W., 1882, Bull. Mus. Comp. Zoél. at Harvard Coll., Vol. IX. p. 300, Fig. 24, Pl. I. Specific Characters.— The bell is hemispherical, with a slight apical pro- jection, and is 4 mm. in diameter. There are 12 long tentacles with well- developed basal bulbs. Each tentacle arises from the point of juncture ofa radial tube with the circular vessel. The velum is well developed. 4 radial vessels arise from the proboscis, and each of these gives rise to two side branches, so that 12 radial tubes reach the circular vessel. In addition to these 12, very slender tubes branch off at right angles to the cireular vessel and end blindly in the gelatinous substance of the bell. These tubes alternate with the 12 radial tubes and tentacles. Each one terminates under a cluster of nematocysts upon the outer surface of the bell. The proboscis is long and slender. and reaches about three quarters of the distance from the apex of the bell cavity to the velar opening. It is provided with 4 slightly recurved and fimbricated lips. This species is remarkable in that a stolon arises from each 48 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. of the 4 main radial canals near to their point of juncture with the proboscis. Each of these stolons gives rise to a number of medusa buds. The meduse become free and thus the species is perpetuated. The proboscis is of a decided sage-green, and the entoderm of the basal bulbs of the tentacles is brown. A number of specimens of this medusa were found at the Tortugas, Florida, in June. A single specimen was found by Brooks at Beaufort, North Carolina. Brooks considered it to be an asexual form of Willetta ornata, and this explan- ation may prove to be correct; we have not found the sexual form of W. ornata, however, at the Tortugas, and incline to regard it as a distinct species. The species differs from the common Willia ornata, A. Agassiz, of Buzzard’s Bay and Newport Harbor, in that the proboscis is far more slender, the nar- row tubes branching off from the circular vessel end each in a single cluster of nematocysts, instead of several clusters as in Willetta ornata ; and above all, the possession of stolons bearing medusa buds separates this form from all other known Atlantic species of Willetta. It is interesting to notice that Huxley (1891, Anatomy Invert. Anim., p. 120, Figure 17) took a species of Willsia (Willetta) in the north Pacific, in which medusa-bearing stolons were developed at the point of bifurcation of each of the four main radial canals. LAODICBEA, Lessoy, 1848. Laodicea neptuna, nov. sp. Figs. 50-52, Plate 20. Specific Characters.— The bell is a little more than a hemisphere, and is 2.5 mm. in diameter. There are 8 short tentacles with large basal bulbs, and 8 small rudimentary tentacle bulbs. The tentacles are thickly covered with nematocysts and are usually carried coiled in a contracted bunch. A single, large, black ocellus is found at the base of each tentacle. There are numer- ous small nematocyst-bearing cirri upon the bell margin between the tentacles. The velum is well developed. There are four straight radial tubes, the upper regions of which, adjacent to the proboscis, are occupied by the gonads. The proboscis reaches slightly beyond the velar opening, and the lips are sur- rounded by 4 prominent clusters of nematocyst cells. The color of the entoderm of the proboscis, tentacle bulbs, and circular and radial tubes is pearly-white. The entodermal lamella of the bell is of a delicate shade of ereen. This medusa was occasionally found at the Tortugas, Florida, during July and August, 1898. MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 49 Laodicea ulothrix, Harcket. Laodicea ulothrix, Haeckel, E., 1879, Syst. der Medusen, p. 133, Taf. VIIL., Figs. 5-7. Specific Characters. — The bell is about 20 mm. in diameter and is about twice as broad as it is high. (Haeckel, 1879, p. 133, says “etwa doppelt so hoch als breit.’”?) This is doubtless a misprint. There are 70-100 long, slender, stiff tentacles, the distal ends of which are coiled in a close helix. The basal bulbs of these tentacles are large and hollow, and there is a well- developed ectodermal ocellus upon the inner (centripetal) side of each bulb. In addition to these ocelli one often sees small spur-like projections upon the outer (centrifugal) sides of the tentacle bulbs. Not all of the tentacles pos- sess these spurs. Sensory clubs and cirri are scattered somewhat irregularly between the tentacles. The sensory clubs are almost as numerous as the tentacles. They are flask-shaped, and their entodermal cores are in direct connection with the entoderm of the circular tube. There are no otoliths. The cirri are usually less numerous than the tentacles. They are coiled ina helix, and their distal ends are covered with large spindle-shaped nematocyst- capsules. The velum is well developed. There are 4 straight, narrow, radial tubes, the proximal halves of which, adjacent to the proboscis, are occupied by the gonads. The proboscis is short, and there are 4 recurved lips. The entoderm of the proboscis, gonads, and tentacle bulbs is brownish- white, or greenish-white in color. This medusa is one of the commonest at the Tortugas, Florida. Haeckel found it at the Canaries, and Brooks describes it from the Bahama Islands. The distribution of the sensory clubs is usually more irregular than is described by Brooks. TIAROPSIS, Aeassiz, L., 1849. Tiaropsis punctata, nov. sp. Figs. 60-63, Plate 22. Tiaropsis diademata, Fewkes, J. W., 1882, Bull. Mus. Comp. Zodl., Vol. IX. p. 277, Pl. VII. Figs. 13-14. Specific Characters. — The bell is bluntly cone-shaped and is 4 mm. in di- ameter. There are 4 well-developed, radially placed tentacles, the distal ends of which are usually coiled in a close helix. In addition to these there are 4 rudimentary tentacle bulbs. The 8 marginal sense-organs are situated midway between the 8 tentacles. Each of these organs consists of a pocket-like fold of the velum containing 8-13 otoliths. Immediately above the otocyst there is a well-developed, deeply pigmented eye (see Linko, A., 1899; Travaux Soc. Imp. 50 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. des Nat. de St. Pétersbourg, T. XXIX. p. 155, Plate I. Figure 5). The velum is very well developed. There are 4 straight, narrow, radial tubes, upon the upper regions of which the gonads are situated. The proboscis is wide and flask-shaped, and the mouth is provided with four prominent, crenulated lips. The color of the entoderm of the proboscis and tentacle bulbs is ochre- yellow, or reddish-brown. Several specimens were found at the Tortugas, Florida, late in June and early in July, 1898, and in June, 1899. It is evident that this species has been noticed by Fewkes, 1882, under the name of “ Tiaropsis diademata.” The species is quite distinct from T. diade- mata, however, for it is smaller, possesses fewer tentacles, and is of a different color; moreover the bell of the young medusa is very much flatter than is that of T. diademata in a corresponding stage of development. The Tortugas form is closely allied to T. rosex of the Fiji Islands ; and it also bears some resem- blance to T. mediterranea, Metschnikoff (1886; Arbeit Zool. Inst. Wien. Bd. VI. p. 239, Taf. I. Figs. 6-8). OCHANIA, Peron and Lesvenur, 1809. Oceania McCradyi. Figs. 56-59, Plate 21. Epenthesis McCradyi, Brooks, W. K., 1888, Studies Johns Hopkins Univ. Biol. Lab., Vol. IV. pp. 147-162, Pls. 13-15. We present some colored figures of this remarkable medusa which develops hydroid-blastostyles upon its gonads. It has been found by Brooks among the Bahama Islands, and by Bigelow off the Florida Coast. We found it at the Tortugas, Florida, in July, 1898. Brooks, 1888, claims to have found the hydroid of this species. Oceania magnifica, nov. sp. Figs, 18, 182, Plate 9. Specific Characters, —The bell is thin and flat and 14 mm. in diameter. There are 32 slender tentacles of short length. There are 64 otocysts, 2 between each successive pair of tentacles. Each otocyst contains a single, spherical otolith. The velum is small. There are 4 straight, narrow, radial tubes. The gonads are developed upon the distal portion of these tubes near to the circular canal. The proboscis is short, and there are 4 sharply curled lips. The color of the entoderm of the proboscis and tentacle bulbs is intense green, while the ectoderm of the proboscis and of the genital organs is usually rich purple. Several specimens were found at the Tortugas, Florida, in June, 1897, and a large number during the summers of 1898 and 1899. MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. bil Oceania globosa, nov. sp. Figs. 19, 19%, Plate 9. Specific Characters. — The bell is globular in form, 14 mm. in diameter, The cavity of the bell is shallow so that the gelatinous substance is very thick. There are 32 large tentacles and 32 rudimentary ones. There are 64 otocysts alternating with the tentacles. Each otocyst contains 3-5 spherical otoliths (Figure 19*). There are 4 straight, narrow, radial canals. The 4 gonads are situated upon the distal portions of the canals. The proboscis is very short and there are 4 prominent lips. The color of the entoderm of the proboscis and tentacle bulbs is light drab. Single specimen found at Tortugas, June 16, 1897. Oceania gelatinosa, nov. sp. Figs. 20, 20°, Plate 10. Specific Characters.— The bell is 7 mm. high and 3.3 mm. in diameter. The gelatinous substance of the upper portion of the bell is very thick. There are 16 well-developed tentacles and 16 rudimentary ones that may develop later. There are 32 otocysts alternating with the tentacles. Each of these otocysts contains 3-5 spherical otoliths (Figure 20). The velum is promi- nent. There are 4 radial canals, in the upper or proximal portion of which the gonads are developed. The proboscis is long and slender and there are 4 prominent lips. The color of the entoderm of the proboscis and tentacle bulbs is light drab, or opaque white. A specimen was found at the Tortugas, Florida, on June 14, 1897, and several others during the summer of 1899. Oceania discoida, nov. sp. Figs. 53-55, Plate 20. Specific Characters. — The bell is quite flat, with conically sloping sides, and is 4mm. in diameter. There are 16 short marginal tentacles with large basal bulbs. There are usually 3 otocysts between each successive pair of tentacles (see Figure 55). The velum is well developed. There are 4 straight radial tubes, upon the greater portion of the length of which the gonads are situated. Tn the case of the female the eggs are very large and prominent. The probos- cis is urn-shaped and there are 4 recurved lips. The proboscis, gonads, and tentacle bulbs are yellow, or yellow-green. The entodermal supporting lamella of the bell is often of a delicate shade of green. The medusa is easily distinguished from the other species of Oceania at the Tortugas by the circum- stance that it is very small in size, the bell is conical in shape, and the gonads are large'and prominent. It is quite common throughout the summer, 52 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Obelia, sp. Eucope, sp. Agassiz, A., 1881, Bull. Mus. Comp. Zool. at Harvard Coll., Vol. IX. p- 149. A very few specimens of the medusa of an Obelia were found at the Tor- tugas, Florida, late in June, 1898. The bell was disk-shaped and about 3 mm. in diameter. There were 96 tentacles and § otocysts. The gonads were large and spherical, and much distended with ova. The bell was colorless. Not having seen the hydroid stock we are unable to determine whether or not this Obelia is identical with any of the forms found at Newport, R. I. EPENTHESIS, McCrapy, 1857. Epenthesis folleata, McCrapy. Fig. 139, Plate 41. Epenthesis folleata, McCrady, J., 1857, Gymn. Charleston Harbor, p. 89. Oceania folleata, Agassiz, A., 1865, North Amer. Acal., p. 70. Specific Characters. — Adult medusa. The bell is usually flatter than a hemisphere, and is about 5 mm. in diameter. Its cavity is shallow, and the bell walls diminish in thickness very gradually from the summit towards the margin. There are 16 slender tentacles with well-developed basal bulbs. 16 otocysts alternate in position with the tentacles. Each otocyst contains a single spherical otolith. The velum is well developed. There are 4 slender, straight, radial canals and a narrow ring-canal. The proboscis is short and simple and there are 4 slightly recurved lips. The 4 gonads are short and linear and are developed upon the 4 radial canals near the circular canal. In the young medusa they are found higher up upon the radial canals, but they migrate centrifugally as development proceeds. The entoderm of the probos- cis, tentacle bulbs, and gonads is of a decided green color. This medusa is very common at the Tortugas, Florida, during the spring months, but becomes rare after the first of June. We have found it abundant in the Bahamas during the winter months. It is rare at Charleston, South Carolina, but Brooks found both hydroid and medusa at Beaufort, North Carolina. EUCOPIUM, Haecket, 1879. Eucopium parvigastrum, nov. sp. Fig. 140, Plate 42. Specific Characters. — The bell is half egg-shaped and is 1 mm. in height. There is a very small apical projection. There are 4 very small radially MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 53 situated tentacles, which are hardly more than mere tentacle bulbs. There are 8 otocysts, 2 in each quadrant. Each otocyst contains a single spherical otolith. The velum is well developed. There are 4 straight, narrow, radial canals, and a slender circular vessel. The proboscis is very small, and is a mere tube, cruciform in cross-section and provided with 4 simple lips. The gonads occupy 4 linear swollen regions near the mid-regions of the 4 radial canals. The entoderm of the tentacle bulbs, gonads, and proboscis is of a decided brown color. This medusa was quite common at the Tortugas, Florida, late in June, 1899. The very small proboscis and marginal tentacles as well as the remarkable swollen condition of the gonads in this medusa foreshadow the condition of Agastra mira (Hartlaub, 1897 ; Wissen. Meeresuntersuch. Biol. Anstalt Hel- goland, Neue Folge, Bd. II. p. 504, Tat. XII. Fig. 10), where there is no trace either of proboscis or tentacles. PSEUDOCLYTIA, nov. gen. Pseudoclytia pentata, nov. sp. Figs. 24-26, Plate 12; Figs. 35, 354, Plate 15: Figs. 131, 132, Plate 39. Generic Characters. — Pseudoclytia. Eucopide with numerous simple ten- tacles (20 in this species). Otocysts alternating with the equally numerous tentacles. 5 simple radial canals, 72° apart. 5 gonads situated upon the 5 radial canals. The proboscis lacks a peduncle and is provided with 5 simple lips. Specific Characters. — Adult medusa. The bell is flatter than a hemisphere and is 8-13 mm. in diameter. There are 20 simple tentacles with well-de- veloped basal bulbs. Each of these tentacles is a little less than half as long as the bell height. There are no lateral or marginal cirri. There are 20 otocysts which alternate in position with the 20 tentacles. Each otocyst con- tains a single spherical otolith (Figure 26). The velum is well developed. There are 5 straight, narrow, radial canals 72° apart. The 5 gonads are situ- ated upon the radial canals at points midway between the proboscis and the bell margin (Figures 35, 131). In the female the ova are large and prominent, and when immature are seen to have a well-defined nucleus and nucleolus (Figures 35*, 131). The proboscis is flask-shaped and there are 5 simple recurved lips. The entoderm of the proboscis, gonads, and tentacle bulbs is usually slightly milky in color, with a few scattered cinnamon-colored granules. Occasionally an individual is met with in which these cinnamon- colored granules are developed to such an extent that the medusa displays a brick-red color (Figure 35). In most individuals, however, the colored gran- ules are so faint as to be almost imperceptible. In some individuals there is a more or less decided green spot in the entoderm of each tentacle bulb (Fig- ures 131, 132). 54 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. This medusa is very common throughout the summer months at the Tortu- gas, Florida. On July 22, 1898, a great swarm of them appeared, and were so abundant that one could not dip up a bucketful of sea-water without captur- ing several specimens; and two such swarms came in the summer of 1899, This is the only Hydromedusa known which is normally formed upon the plan of five (pentamerous). It seems very probable that it has arisen, phylo- genetically, as a sport from some species of Epenthesis or Oceania, some indi- viduals of which made their appearance with 5 radial canals instead of 4; and these abnormal individuals succeeded in perpetuating a new species. Bateson (1894; Materials for the Study of Variation, p. 425) calls attention to an ab- normal specimen of Sarsia mirabilis having five complete segments, and says that “there is perhaps in the whole range of natural history no more striking case of the Discontinuity and perfection of Meristic Variation. In the case of Eucope (Obelia) it has been shown by Agassiz and Woodworth (1896; Bull. Mus. Comp. Zodl. at Harvard Coll., Vol. XXX. p. 121-150, 9 Plates) that among 3,917 meduse 9 had three radial canals, 20 had five, and 3 had six radial canals. It thus appears that in Obelia the tendency to produce sports having 5 radial canals is about twice as great as that to produce individuals with any other number of canals. Yet sports of Obelia with 5 radial canals have not succeeded in perpetuating a new species. I have made careful observations of 1000 individuals of Pseudoclytia pen- tata, and find that 70.3% are normal (7. e. have 5 canals 72° apart, 5 gonads, and 5 lips to the proboscis). The remaining 29.79% are abnormal in some respects, and a large number of the abnormalities tend toward the ancestral condition of 4 canals and 4 lips. The medusa is very much more variable than the 4-rayed Epenthesis folleata at the Tortugas, and its greater variability may be due to the fact that being a new form it displays a greater tendency toward variability in various directions. This question will, however, be made the subject of a special paper. - MULTIORALIS, nov. gen. Multioralis ovalis, nov. sp. Figs. 129, 130, Plate 39. Generic Characters. — Multioralis. Leptomeduse having a circular canal, and a single, simple chymiferous canal which extends across the sub-umbrella. A number of separate manubria are situated upon the chymiferous canal. Specific Characters. — Adult medusa. The bell is quite flat, and is ellipti- cal in outline, the major axis being 4 mm. and the minor 2.4mm. The gelati- nous substance is not very thick and is quite flexible. There are 20-25 short, simple, coiled tentacles with well-developed basal bulbs. These tentacles are only about one half as long as the minor axis of the bell. There are no lateral or marginal cirri. The otocysts are slightly more numerous than the tenta- MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 55 cles; usually one, but occasionally two, being found between each successive pair of tentacles. Hach otocyst contains a single spherical otolith. The velum is simple and quite broad. There is a slender circular vessel, and a single straight chymiferous canal extends along the major axis of the bell. In the oldest medusz observed there were 4 manubria. ‘Two equally developed large manubria were situated on either side of the centre of the sub-umbrella, upon the chymiferous canal; while two small manubria were found upon the same canal centrifugally away from the larger manubria. There was thus no manubrium at the centre of the sub-umbrella. There were two small gonads upon the chymiferous canal immediately centrifugal from the small manu- bria. The entoderm of the manubria and of the basal bulbs of the tentacles is of an opaque glistening white. The supporting lamella of the bell is of a deli- cate green. Young Medusa.—In the youngest medusa observed, there were but 2 manubria situated upon the chymiferous canal on either side of the centre of the disk. The major axis of the bell was 2.5 mm. and there was no trace of gonads. About a dozen specimens of this medusa were captured at the Tortu- gas, Florida, from June 30-July 2, 1899. It seems possible that the bell of the large medusze may divide by transverse fission, for one individual was found in which there was a decided notch in the bell-margin extending inward in the plane passing through the centre of the sub-umbrella perpendicular to the main chymiferous tube. This notch ap- peared, however, upon only one side of the bell and may have been due to an accident. The main chymiferous canal is of course equivalent, morphologi- cally, to two diametrically opposed radial canals. HUCHEILOTA, McCrapy, 1857. Eucheilota ventricularis, McCrapy. Fig. 128, Plate 38. Eucheilota ventricularis, McCrady, J., 1857, Gymn. Charleston Harbor, p. 85, Pl. 11, Figs. 1,2; Pl. 12, Figs. 1-3. This medusa is quite rare at the Tortugas, Florida, and not more than a dozen specimens were obtained. They were remarkable in that the entoderm of the tentacle bulbs and proboscis was of a decided green color. Hach otocyst contained 2-4 spherical otoliths. In specimens 2 mm. in diameter there were as yet no gonads upon the radial canals. 56 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. Eucheilota bermudensis. Oceanopsis bermudensis, Fewkes, J. W., 1888, Bull. Mus. Comp. Zodl., Vol. XI. p. 86, Pl. I. Figs. 8-10. Specific Characters. — Adult medusa. The bell is not quite hemispherical, the sides being relatively straight and sloping and the top quite flat. It is about 6 mm. in diameter. There are 8 tentacles, 4 radial and 4 interradial. These tentacles are only about one third as long as the bell diameter, and are thickly covered with nematocysts. Their basal bulbs are large, and are each flanked by a pair of short, coiled, nematocyst-bearing cirri. In addition to these there are normally about 8 other cirri in each quadrant, and scattered between them are 8 otocysts. Thus the medusa has 32 otocysts and 48 cirri. Kach otocyst is of small size and contains a single spherical otolith. The velum is well developed. There are 4 straight, narrow, radial canals, and a simple circular vessel. The proboscis is short, but wide, and there are 4 cruciform, slightly recurved lips. There is no peduncle. The 4 gonads are found upon the 4 radial canals near the proboscis. These are visible in young meduse about 1.5 mm. in diameter; and in the adult they become quite large and swollen, the ova being distinctly seen lying along the side of the canal. The entoderm of the proboscis gonads and tentacle bulbs is grass green, and the supporting lamella of the bell is tinged with the same color. This medusa was quite common at the Tortugas, Florida, from June 17-25, 1899. Young Medusa.—The youngest medusa was about 2 mm. in height and 1.5 mm. in diameter. It was very much in the condition described by Fewkes, 1883. There were 4 simple radially situated tentacles and 4 interradial ten- tacle bulbs. The interradial tentacle bulbs were flanked by lateral cirri, while the radial tentacle bulbs lacked these appendages. There were 4 otocysts, one upon the side of each of the interradial tentacle bulbs. Each otocyst contained a single spherical otolith. The gonads were already quite large, and lay along the 4 radial canals near the sides of the proboscis. The proboscis was short, and there were 4 simple lips. Eucheilota paradoxica, nov. sp. Figs. 134-136, Plate 40. Specific Characters. — Adult medusa: the bell is somewhat fuller than a hemisphere and is 4 mm. in diameter. The gelatinous substance is of moderate thickness, and there is a very slight, blunt, aboral projection. There are 4 equally developed, radially situated tentacles. These tentacles are about as long as the bell height, but are usually carried coiled in a close helix. Their basal bulbs are elongate, and are hollow. The shafts of these tentacles are thickly covered with nematocysts. A pair of tightly coiled lateral cirri arise from the sides of each tentacle bulb. In addition to these well-developed tentacles there MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 57 are 4 interradial, rudimentary tentacle bulbs which are flanked by lateral cirri. There are 8 otocysts, 2 in each quadrant. Each otocyst contains a single spherical otolith. The velum is well developed. There is a narrow circular vessel, and 4 straight simple radial canals. The proboscis is flask-shaped, and there is no peduncle. There are 4 simple cruciform lips. Medusa buds in various stages of development are found upon the 4 gonads, which are situated at the middle points of the 4 radial canals. These medusa buds first develop 2 diametrically opposed tentacles (Figure 135), but when about to be set free they have 4 equally developed tentacles as in the adult. ‘They have, however, no trace of gonads, and the interradial tentacle bulbs are not provided with lateral cirri. Usually from 2-5 medusa buds in several stages are found upon each gonad. The entoderm of the proboscis gonads and tentacle bulbs is of a milky-green color. This medusa was common at the Tortugas, Florida, in June, 1899. This is the first and only Leptomedusa which has been observed to give rise to free medusa buds. EUTIMA, McCrapy, 1857. Eutima mira, McCrapy. Eutima mira, McCrady, J., 1857, Gymn. Charleston Harbor, p. 88, Pl. XI. Figs. 8, 9. This medusa is common throughout the summer at the Tortugas, Florida. It is also abundant at Charleston, South Carolina, and at Beaufort, North Caro- lina. Damaged specimens are occasionally drifted into Newport Harbor, Rhode Island, by the southerly winds, late in the summer. EUTIMALPHHES, Haecxet, 1879. Eutimalphes coerulea. Figs. 22, 22°, Plate 11. Eirene cerulea, Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV., p. 362. Irene cerulea, Haeckel, E., 1879, Syst. der Medusen, p. 203. Specific Characters. — The bell is 10 mm. in diameter and a little broader than it is high. The gelatinous substance at the apex of the bell is quite thick, but becomes progressively thinner as one approaches the margin. There are about 32 short, slender, marginal tentacles, each one of which is furnished with small lateral cirri. In addition to the well-developed tentacles there are about 96 rudimentary tentacular swellings upon the bell margin. There are usually about three of these swellings between each successive pair of tenta- cles. (Figure 22%.) There are 8 otocysts, 2 in each quadrant, and each one of them contains 3-5 spherical otoliths. There are 4 radial tubes. The 58 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. velum is well developed. There are 4 linear, slightly convoluted gonads. They begin about halfway between the circular vessel and the peduncle, and extend to a point close to the proboscis. The peduncle is well developed and reaches slightly beyond the velar opening. The gastric portion of the pro- boscis is short and is furnished with 4 slightly fimbricated lips. The probos- cis, gonads, and tentacle bulbs are opaque white. Common at the Bahamas and Tortugas in the spring and winter months. EUTIMIUM, Hasrcxez, 1879. Eutimium serpentinum, nov. sp. Figs. 69-72, Plate 23. Specific Characters. — The bell is 10 mm. in diameter, and about 24 times as broad as it is high. There are 4 radially situated tentacles; each being about as long as the bell diameter. There are no lateral or marginal cirri. The 8 otocysts are situated near to and on both sides of the radial tentacles. (Figure 70.) Each otocyst contains 4-8 spherical otoliths. The velum is well developed. There are 4 straight narrow radial tubes and a narrow circular vessel. The proboscis possesses a very long peduncle, which is about 3 times as long as the bell diameter. The upper region of the peduncle is conical in shape; then follows a long slender cylindrical region leading to the gastric part of the proboscis, which is urn-shaped with 4 slightly recurved lips. (Figure 71.) The 4 gonads are situated upon the long cylindrical portion of the peduncle, where they lie upon the radial canals. (Figure 72.) The pro- boscis, gonads, and tentacles are opaque bluish-white. Half a dozen specimens of this medusa were found at the Tortugas, Florida, late in July, 1898. It is closely allied to Eutimium elephas, Haeckel (1879 ; Syst. der Medusen, p. 190, Taf. XII. Figures 10-12), of the German Ocean. PHORTIS, McCrapy, 1857. Phortis lactea, nov. sp. Fig. 133, Plate 40. Specific Characters. —The bell is 5 mm. in diameter and the sides flange slightly outward at the margin. The gelatinous substance is of moderate thickness at the aboral pole, but becomes thin at the margin of the bell. There are about 18-22 short simple tentacles, the basal bulbs of which are large and swollen. These tentacles are only about one fifth as long as the bell diameter. There are no lateral or marginal cirri. The otocysts are slightly more numer- ous than the tentacles, there being at least one, and occasionally two, of these structures between each successive pair of tentacles. Each otocyst contains a MAYER: MEDUS FROM THE TORTUGAS, FLORIDA. 59 single spherical otolith. The velum is well developed. There are 4 straight slender radial canals, which extend down the peduncle to the gastric portion of the proboscis. The peduncle is wide at its base, but not so wide as in Phortis pyramidalis. It extends for a short distance beyond the velar opening of the bell. The gastric portion of the proboscis is cruciform in cross-section and there are 4 simple recurved lips. The 4 gonads are situated upon the 4 radial canals a short distance above their junction with the circular vessel. Each gonad is linear, and in the female the ova are quite conspicuous. The gonads and the gastric portion of the proboscis are milky in color, while the tentacle bulbs are cream-colored with greenish entodermal granules, Found at the Tortugas, Florida, in June. Phortis pyramidalis. Figs. 21, 212, Plate 10. Eutima pyramidalis, Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV., p. 363. Specific Characters. — Adult medusa. The bell is slightly flatter than a hemisphere, and attains a diameter of about 35 mm. There are about 100 small slender tentacles, which lack lateral cirri. About 100 otocysts alternate with the equally numerous tentacles. Each otocyst contains a single spherical otolith. (Figure 21%.) There are 4 narrow radial canals. The proboscis is provided with a wide cone-shaped proboscis which fills most of the cavity of the bell, and projects outward for a considerable distance beyond the velar opening. The gastric portion of the proboscis is very small, and is provided with 4 delicately crenulated lips. The gonads are linear and are developed upon the centrifugal portions of the 4 radial canals near to the circular canal. The proboscis, tentacle bulbs, and gonads are of a delicate blue-green color. This medusa is very abundant among the Bahama and Tortugas Islands. At night, when disturbed, it glows with an intense blue-green phosphorescence which is far more brilliant than that of any other medusa that we have observed. Young Medusa. — Phortis pyramidalis. In the youngest medusa observed the bell was higher than a hemisphere and 3 mm. in diameter. There was no peduncle to the proboscis, and the gelatinous substance of the bell was not very thick. There were 4 slender radial tubes and 16 tentacles, only 8 of which had attained to any length, the others being mere basal bulbs, There were about 8 otocysts, each containing a single spherical otolith. When the medusa is about 7 mm. in diameter, the bell is flatter than a hemisphere. The peduncle is well developed and extends beyond the velar opening. The gastric portion of the proboscis has grown very little and is relatively to size of the medusa much smaller than in the younger animal. There are 4 re- curved lips. There are now about 32 tentacles and 16 otocysts. VOL. XXXVII. — NO. 2. 5 60 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. ZYGODACTYLA, Branpr, 1835. Zygodactyla cubana, nov. sp. Figs. 84, 85, Plate 25. Specific Characters. — Young medusa. The bell was quite flat and disk- shaped and 4.5 mm. in diameter. There were 8 long tentacles, 8 rudimentary, undeveloped tentacles, and 16 very small undeveloped tentacle bulbs, that probably develop later into tentacles. The tentacle bulbs possessed excretion papille and were further distinguished by the fact that there were two ento- dermal green pigment spots one on either side of the bulb (see Figure 85). These spots had the appearance of ocelli, but we do not venture to state that they are such. There were 32 otocysts, each containing one or two spherical otoliths. The velum was well developed. There were 16 radial canals, only 8 of which reached the circular vessel. The 8 others projected about half- way from the proboscis to the circular canal. The proboscis was wide and flask-shaped, and projected for a considerable distance beyond the velar open- ing. The 16 lips were recurved. The gonads were beginning to appear upon the radial canals. The entoderm of the proboscis and radial canals is sage- green. The entoderm of the tentacle bulbs was flesh-colored and the “ ocelli” were green. Tortugas, Florida, July 25-29, 1898, and June, 1899. Zygodactyla cyanea, Acassiz, L. Figs. 23, 234, Pl. 11; Figs. 33, 34, Pl. 15. Zygodactyla cyanea, Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV. p. 361. Mesonema cyaneum, Haeckel, E., 1879, Syst. der Medusen, p. 227. Specific Characters. — Adult medusa. None of our figures were drawn from full-grown meduse. The bell is flatter than a hemisphere and is about 45 mm. in diameter (22 mm. in Figure 33). The gelatinous substance of the central part of the bell is very thick and there is a well-developed peduncle which projects downward into the cavity of the stomach. The peripheral zone of the bell is quite thin and flexible. There are 90-100 well-developed tentacles with large conical basal bulbs. Each tentacle bulb is hollow and is provided with a conical excretion papilla which projects outward (centrifugally). See Figure 34, Plate 15. There are one or two (usually one) otocysts between each suc- cessive pair of tentacles. Each otocyst contains one or two spherical otoliths. The velum is well developed. There are 90-100 simple, straight radial tubes, The radial tubes do not extend down the peduncle of the proboscis, but empty into the stomach cavity at their highest point. The proboscis is wide and shal- low, and does not protrude beyond the velar opening. The mouth is sur- rounded by numerous crenulated lips which are equal in number to the radial MAYER: MEDUS FROM THE TORTUGAS, FLORIDA. 61 canals, The stomach is about two thirds as wide as the bell diameter. The gonads are linear, and occupy almost the whole length of the radial canals. The entoderm of the gonads, tentacle bulbs, and proboscis is blue-green. The medusa is very common off the Florida Coast both in summer and winter. AEQUORBA, Péron and Lesveur, 1809. Afquorea floridana. Rhegmatodes floridanus, Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV. p. 361. Specific Characters. — The bell is hemispherical and 25 mm. in diameter. The gelatinous substance of the bell is thick and of a tough consistency. There are 16 radial tubes and 64 well-developed marginal tentacles. The tentacle bulbs are large and hollow, and are provided each with one or two excretion papille, which project outward from the side of the bell. There are about 192 otocysts, three between each successive pair of tentacles. Each otocyst contains 2 spherical otoliths. The velum is well developed. The 16 gonads are developed upon the distal halves of the 16 radial canals. The sur- face of the mature gonads is slightly convoluted. The proboscis is wide and very shallow, and there are 16 slightly fimbricated lips. The gonads and the edge of the bell are milky-white. Common at the Tortugas and Bahamas in the spring months. RHACOSTOMA, Agassiz, L., 1862. Rhacostoma dispar, nov. sp. Figs. 27-29, Pl. 13. Specific Characters, — The bell is lens-shaped, and about 40 mm. broad and 20 mm. high. The cavity of the bell is remarkably small and shallow, so that the gelatinous substance is very thick. The velar opening of the bell is only about 5 mm. in diameter. There are about 8 very small rudimentary tentacles. There are 30-40 otocysts scattered between the tentacles. Each otocyst con- tains 3-5 oval-shaped otoliths (see Figure 29). There are no excretion papille. The velum is well developed. There are about 80 radial tubes, fully half of which end blindly without reaching the circular tube. The gonads are situated upon all of the radial tubes. They are linear, and their surfaces are slightly convoluted. They do not extend quite to the peripheri of the stomach, nor do they reach the circular canal. The proboscis is very wide, and may at times be protruded beyond the velar opening. There are about 80 small crenulated lips, which are apparently as numerous as the radial canals. The bell has a faint steel-blue tinge, and the genital organs are pink. A single specimen 62 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. was found at the Tortugas, Florida, in June, 1897. This remarkable species is extremely inactive. Owing to the small size of the velum, it is of but little service in swimming, and the medusa makes use of the contractions of its widely open mouth in order to propel itself through the water. GONIONEMUS, Agassiz, A., 1865. Gonionemus aphrodite. Cubaia aphrodite, Mayer, A. G., 1894, Bull. Mus. Comp. Zodl. at Harvard Coll., Vol. XXV. p. 237, Pl. II. Figs. 1-3. This medusa is occasionally met with at the Tortugas, Florida, and an examination of mature individuals has convinced me that it belongs to the genus Gonionemus. The gonads consist of a series of finger-shaped, or papilli- form, processes that are crowded alternately to one side and the other of the radial canal very much as in the species of Gonionemus found at Woods Holl, Massachusetts. GONIONEMOIDES, nov. gen. Gonionemoides geophila, nov. sp. Figs. 6-11, Plates 3-5. Generic Characters. — This genus is closely related to Gonionemus, but differs from it in that the marginal tentacles are of two distinct kinds, and arise at slightly different levels from the bell margin. One of these sets of tentacles is provided with nettling cells, and the other is furnished with adhesive suckers, as in Gonionemus. There are 4 radial canals, and the circular vessel is simple without centripetal canals. The gonads are papilliform and are situated upon the radial canals. There ate numerous otocysts upon the bell margin. Specific Characters. — Adult medusa, Figures 6-9. The bell is quite flat and disk-shaped, and is about 9.5 mm. in diameter. There are 64 marginal tentacles. 16 of these bear, each one, a suctorial disk upon the aboral sides near their distal extremities. The extreme distal ends of the tentacles are cone-shaped, and are bent sharply at a right angle to the main shaft of the tentacle, very much as is the casein Gonionemus vertens, A. Agassiz. These sucker-bearing tentacles arise at a level, a little above the bell margin. The remaining 48 tentacles all arise from the bell margin, at a lower level than do the sucker-bearing ones. They possess no suctorial disks, but instead are armed with rings of nematocyst capsules (Figure 6). These nematocyst-bear- ing tentacles are far more flexible than are the sucker-bearing ones. There are 12 otocysts upon the bell margin, each one of which contains a single otolith situated within an elongate, oval cavity (see Figure 7). The MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 63 velum is well developed. There are 4 straight radial tubes. The gonads occupy the distal halves of the radia] tubes, but do not quite extend to the circular vessel. ‘They present the appearance of a series of papilliform, or fin- ger-shaped, processes that are crowded alternately to one side and the other of the radial tube, very much as is the case in Gonionemus vertens. The probos- cis is a simple tube with 4 prominent lips. There are 4 radially situated green-colored spots upon the proboscis close to its junction with the 4 radial canals. Young Medusa. — Figures 10, 11, Plate 5. The youngest medusa observed was 1.7 mm. in diameter. The bell was high and quite thick, and its aboral surface was covered with nematocyst capsules. The 16 sucker-bearing tenta- cles were already present, although the suctorial disk was visible upon only 8 of them. Figure 11 is a side view of the end of one of these young tentacles showing the beginning of the formation of the suctorial disk. There were 7 otocysts present. The velum was very prominent. There were no traces of genital organs present. The proboscis possessed a distinct peduncle. The color of the genital organs, bell margin, and proboscis of this medusa is pearly white. The entoderm of the tentacle bulbs and of the radial tubes in the region of the gonads is green. The ocelli? of the proboscis are green. The adult medusa would frequently lie flat upon the bottom of the aquarium with its oral surface upward (Figure 9, Plate 4). In this position the sucker- bearing tentacles would be stretched far out and the suckers would anchor the medusa to the bottom. The nematocystic tentacles, on the other hand, would wave freely upward apparently in position for the capture of prey. When disturbed the medusa would swim actively abont for a few moments, and then reassume its characteristic position of rest. This medusa was common at Key West from May 27-June 10, 1897. HALICALYX, Fewxss, 1882. Halicalyx tenuis, FewKes. Figs. 12, 13, Plates 5, 6. Halicalyx tenuis, Fewkes, J. W., 1882, Bull. Mus. Comp. Zo6l. Harvard Coll., Vol. IX. p. 277, Pl. VII. Fig. 15. Generic Characters. — This genus is closely allied to Gonionemoides, but differs from it in that none of the tentacles bear suctorial disks. The tentacles are of two distinct kinds and arise at different levels from the bell margin. The circular vessel gives off blind centripetal branches. There are otocysts at the bases of the tentacles. Tentacles numerous. Specific Characters. —The bell is 25 mm. in diameter and is hemispherical. It is quite thick at the aboral pole, but becomes constantly thinner as one approaches the margin. The gelatinous substance is of very rigid consistency. 64 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. There are about 50 tentacles, and 64 short, blunt papillz upon the bell margin. 32 of the tentacles arise from the side of the bell at a little distance above the margin. They are short and stiff and stand out sharply at right angles to the bell (see Figure 3). These tentacles are sprinkled over with wart-like pro- tuberances of a deep purple color. A pair of otocysts, each containing a single otolith, are situated at the base of each of these stiff tentacles. Thus the medusa possesses 64 otocysts. In addition to the stiff tentacles there are about 20 others that are long and flexible, and arise from the bell margin. They are covered with rings of nematocyst cells closely coiled in a helical manner (see Figure 12). These tentacles are very flexible and are constantly being expanded to a length of 4-5 times the diameter of the bell and then contracted with a sudden jerk. The velum is small. There are 4 straight, narrow, radial tubes. The circular vessel is peculiar in that it gives off blindly ending, cen- tripetal branches or diverticule, that penetrate inward into the substance of the bell. There are in all 28 side branches. 4 of these are each about half as long as the radial tube. 8 others are only one quarter as long as this, and the 16 remaining ones are still shorter. These diverticule are situated immediately above the short, stiff tentacles (see Figure 13). The gonads are found occu- pying the distal half of the radial canals, but do not reach quite to the bell margin. They hang downward into the bell cavity as a complex system of finger-shaped papille. The proboscis is very slender and the lips prominent. It extends for about three quarters of the distance of the height of the bell cavity. The gelatinous substance of the bell is slightly greenish in color. The entoderm of the proboscis, genital organs, circular tube, and tentacles is opaque yellow-green and reddish purple. There are 4 reddish-purple spots upon the proboscis just between the radial canals. This medusa was common at Key West from May 27-June 10, 1897. It was extremely active in all of its movements and wonderfully hardy in cap- tivity. One specimen lived for more than a week in a small glass bowl, the water of which was not changed. It seems probable that both this species and Gonionemoides geophila prefer the muddy and impure waters of the Florida Coast, for while they were both common at Key West, they were not seen at the Tortugas either in 1897, 1898, or 1899. LIRIOPH, Lesson, 18438. Liriope scutigera, McCrapy. Liriope scutigera, McCrady, J., 1857, Gymn. Charleston Harbor, p. 106. Xanthea scutigera, Haeckel, E., 1864, Geryoniden, p. 24. Liriantha scutigera, Haeckel, E., 1879, Syst. der Medusen, p. 287. This medusa is not very common at the Tortugas, Florida. It is quite abundant at Charleston, South Carolina, and we have taken it at various places among the Bahama Islands, and off the Cuban Coast, during the winter months. MAYER: MEDUSA) FROM THE TORTUGAS, FLORIDA. 65 GLOSSOCODON, Haecxet, 1864. Glossocodon tenuirostris, FewKeEs. Figs. 75-78, Plate 24. Liriope tenuirostris, Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV. p. 365. Glossocodon tenuirostris, Fewkes, J. W., 1882, Bull. Mus. Comp. Zodél. at Harvard Coll., Vol. IX. p. 278, Pl. VII. Figs. 1-9. Liriope cerasiformis? Maas, O., 1893, Ergeb. der Plankton Exped., Bd. II. K. C., p. 35, Taf. II. Fig, 5. 6. This medusa is common at the Tortugas, Florida; as indeed it is also among the Bahama Islands and along the Cuban Coast. It is met with in consider- able numbers in Charleston Harbor, South Carolina; and occasionally a damaged individual is drifted into Newport Harbor, Rhode Island, by the southerly winds late in the summer months. AGLAURA, Péron and Lesvevr, 1809. Aglaura hemistoma, Péron and Lesvevr. Figs. 79, 80, Plate 25. Aglaura hemistoma, Péron, F., et Lesueur, C. A., 1809, Tableau des Méduses, p. 351, No. 73. Aglaura Peéronii, Leuckart, R., 1856, Archiv fiir Naturges. Jahrg. 22, p. 10, Taf. I. Figs 5-7. This medusa is occasionally met with in June at the Tortugas, Florida. It is found also in the Mediterranean and is widely distributed throughout the tropical regions of the Atlantic (see Maas, O., 1893, Die Craspedoten Medu- sen der Plankton Expedition, Taf. VII.). A very closely allied species is found in the Tropical Pacific (see Agassiz and Mayer, 1898 ; Bull. Mus. Comp. Zodl. at Harvard Coll., Vol. XXXII. p. 166). One figure is drawn from a specimen obtained at the Tortugas, Florida. Aglaura hemistoma, var. Nausicaa, Harcket. Aglaura Nausicaa, Haeckel, E., 1879, Syst. der Medusen, p. 274, Taf. XVI. Fig. 1. Aglaura vitrea, Fewkes, J. W., 1882, Bull. Mus. Comp. Zoél. at Harvard Coll., Vol. IX. p. 277, Pl. VU. Fig. 10. Aglaura hemistoma, var. Nausicaa, Maas, O., 1892, Die Craspedoten Medusen der Plankton Expedition, Bd. II. K. C., p. 26. This variety is occasionally met with at the Tortugas, Florida. (op) jor) BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. CUNOCTANTHA, Haecxet, 1879. Cunoctantha incisa, nov. sp. z Figs. 145, 146, Plate 44. Specific Characters. — The bell is slightly flatter than a hemisphere and is about 5 mm. in diameter. There is a slight apical projection, which is solid. 8 stiff tentacles arise from the sides of the bell, about halfway between the mar- gin and the apex. These tentacles are provided with well-developed conical in- sertions, and their entodermal cells are disk-shaped and highly vacuolated. There is a well-developed peronium beneath each tentacle. The tentacles are all of equal length, and are about three quarters as long as the bell diameter. The ex-umbrella extends outwards in 8 lobes, which are held together in a web formed of the ascending velum. These are 24 pear-shaped marginal sense-or- gans, each containing a crystalline otolith and surrounded by a sensory pad bearing delicate bristles. The lower velum is well developed. The proboscis is flat and the stomach cavity is small. The mouth is a simple opening without prominent lips. The stomach gives rise to 8 pouches which extend outward in the radii of the 8 tentacles. The incisions between these pouches are deeper than in Cunoctantha octonaria of Charleston Harbor, The entoderm of the tentacles, and sometimes of the stomach, is green. Two specimens of this medusa were found at the Tortugas, Florida, late in May, 1899. AZBGINELLA, Haecket, 1879. Ajginella dissonema, HaEcKeL. Figs. 30-32, Plate 14. Eginella dissonema, Haeckel, E., 1879, Syst. der Medusen, p. 340, Taf. XX. Fig. 16 Zginella dissonema, Agassiz, A., and Mayer, A. G., 1899, Bull. Mus. Comp. Zodl. at Harvard Coll., Vol. XXXII. p. 166. Specific Characters. — The bell is 3 mm. in diameter, and has the form of a frustum of a cone with rounded apex. It is a little wider than it is high. There are two long tentacles that arise from the sides of the bell at about ¢ of the distance from the margin to the apex. These tentacles are quite stiff and incapable of contraction. They are carried trailing behind the medusa in two straight parallel lines, and are about 3 times as long as the bell height. The entodermal core of each tentacle consists in a row of disk-shaped, highly vae- uolated cells. (See Figure 31.) In addition to the long tentacles there are two very small protuberances (t. Figure 30) that arise from the bell margin, at the foot of the pair of peronial tubes that are situated 90° from the large ten- tacles. Haeckel, 1879, does not mention or figure these protuberances, and it MAYER: MEDUS FROM THE TORTUGAS, FLORIDA. 67 seems probable that they may be absent in some individuals, for we did not observe them in specimens of Aiginella dissonema from the Fiji Islands. In addition to the above-mentioned protuberances there are 4 small interradial swellings situated upon the bell margin. There are 8 sensory clubs, 2 in each quadrant (see Figure 32), each one of which contains a single spherical otolith. The velum is large and powerful and is constantly contracting and expanding with great rapidity. There are 4 peronial double canals, each canal being divided into two by means of a longitudinal septum. The proboscis is small and flat, and the mouth is a simple circular opening. There are 8 interradial pouches that extend outward from the stomach into the substance of the bell. The gonads are developed upon these pouches and in the specimen here fig- ured they contained immature ova. The color of the entoderm of the probos- cis and of portions of the entodermal core of the tentacles is intense golden- green. The gonads in the specimens described by Haeckel were rose-red; in ours they were colorless. A single specimen was found at the Tortugas, Florida, June 19, 1897. This medusa appears to be very widely distributed. Haeckel found it at the Canary Islands, and we found it in Suva Harbor, Fiji Islands, in January, 1898. Il. SCYPHOMEDUS~. NAUSITHOH, Kérriger, 1853. Nausithoé punctata, Ké.urer. Figs. 67, 68, Plate 23; Figs. $7, 88, Plate 26. Nausithoé punctata, Kolliker, A., 1853, Zeit fiir Wissen. Zodl., Bd. IV. p. 323. Nausithoé punctata (Marginata albida), Agassiz, L., 1862, Cont. Nat. Hist. U. S., Vol. IV. pp. 122, 167. Nausithoé albida, Carus, V., 1857, Icones Zodtom., Taf. II., Figs. 17, 22, 28. Specific Characters.— Adult medusa. The umbrella is quite flat and is about 9 mm. in diameter. There are 8 stiff tentacles, each one of which is about # as long as the bell diameter. The main portion of the entodermal core of each tentacle is solid, but as Vanhoffen, 1892, has shown, the basal bulbs of the tentacles are hollow and connected with the adjacent lappet-pouches. There are 8 marginal sense-organs that alternate with the 8 tentacles. As the Hertwigs (1878, Sinnesorgan der Medusen, Figure 2, Plate 9) and Claus (1883, Organ. Entwick. Medusen, Figure 47, Plate 7) have shown, each sense-organ consists of an ectodermal eye, provided with a lens and with nerve fibres; and also of an entodermal otocyst containing a number of otoliths. (See Figure 68, Plate 23.) The 16 marginal lappets are long and flexible, and it is by means of their movements that the medusa is enabled 68 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. to swim through the water with great rapidity. 16 diverticule or pouches from the stomach enter the 16 lappets. Each of these pouches is simple, and does not give rise to any system of canals ramifying through the lappets. The mouth is a simple cruciform opening and there are no oral appendages, or palps. ‘The gastro-vascular cavity is a wide space occupying the whole centre of the umbrella and extending outward into the lappets to form the 16 lappet- pouches. There are 4 groups of gastric cirri, situated in such manner that the 2 diameters passing through them are 45° apart from the 2 diameters passing through the cross formed by the lips of the mouth. All 4 of these diameters pass through the marginal sense-organs. There are about 6 tentacule in each group of gastric cirri, thus making in all about 24 gastric tentacule. The 8 gonads are of entodermal formation, and are found in the 8 tentacular radii. As Claus, 1883, has shown, each one is formed from a pocket-like fold of the entoderm. A band of circular muscles is found in the ectoderm of the sub- umbrella, and radial muscle fibres run out from this band into the 16 marginal lappets. The color of this medusa is quite variable. The gelatinous substance of the bell is usually bluish white or brownish. The gonads are brownish red or, in the case of the males, bright yellow rosin-colored pigment spots are found in the ectoderm of the ex-umbrella, especially upon the lappets. These rosin- colored spots are due to small crystals (see Claus, 1883; Figure 44, Taf. VI.). A young ephyra of this species (see Figures 67, 68; Plate 23) was found by us near Flamingo Key, Bahama Islands, Feb. 9, 1893. It was 2 mm. in diameter. There were as yet no marginal tentacles. The otocysts each contained 5-6 oval otoliths. There were only 4 gastric cirri. A slightly older ephyra has been figured by Claus, 1883; Figure 48, Taf. VII. This medusa is common in the Mediterranean, and is also found among the Bahama and Tortugas Islands. It was described by Vanhoffen, 1893, from near the mouth of the Amazon River. A very closely allied species was found by us in the Fiji Islands, Pacific Ocean. LINERGES, Harcxet, 1880. Linerges mercurius, Harcket. Linerges mercurius, Haeckel, E., 1880, Syst. der Medusen, p. 495, Taf. XXIX. Figs. 4-6. Linerges pegasus ? Haeckel, E., 1880, Syst. der Medusen, p. 495. Vast numbers of ephyre of this medusa are found among the Bahama Islands and along the Florida Coast in March; and the mature meduse are very abundant in June. At times these creatures appear in such numbers that hundreds are captured in every haul of the tow net. They congregate in great windrows, remain abundant for a few days, and then disappear for an indefinite period. MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 69 DACTYLOMBETRA, Aaassiz, L., 1862. Dactylometra lactea, L. AGassiz. Chrysaora lactea, Eschscholtz, F., 1829, Syst. der Acalephen, p. 81, Taf. VII. Fig. 3. Dactylometra lactea, Agassiz, L., 1862, Cont. Nat. Hist. U.S., Vol. 1V. pp. 125, 126; 166. Dactylometra lactea, Agassiz, A., and Mayer, A. G., 1898, Bull. Mus. Comp. Zool. at Harvard Coll., Vol. XXXII. p. 7, Pls. XII., XIII, and Fig. 10, Pl. VII. This medusa is extremely common in Havana Harbor, Cuba, in February, where it swims upon the surface during the afternoon hours. It has been de- seribed by Eschscholtz from the Bay of Rio Janeiro, Brazil. It is occasionally found at the Tortugas, Florida. AURELIA, Péron and Lesvevr, 1809. Aurelia habanensis, nov. sp. Figs. 73, 74, Plate 24; Fig. 86, Plate 26. Specific Characters. — Adult medusa. The bell is 240 mm. in diameter. It is disk-shaped, and the gelatinous substance is quite thick. There are 8 simple marginal lappets, which bear upon their dorsal surfaces, at a slight dis- tance above the bell margin, a row of numerous short tentacles. There are 8 marginal sense-organs that are deeply set within niches situated between the marginal lappets. The radiating chymiferous tubes are very similar to those of Aurelia flavidula, Péron and Lesueur. The mouth-arms, or palps, are long and narrow and extend almost to the bell margin. Their free edges are not lined with a fringe of tentacles as in Aurelia flavidula, but instead are covered with wart-like clusters of nematocyst cells (see Figures 73, 74, Plate 24). This, indeed, constitutes the principal difference between this species and Aurelia flavidula. The 4 gonads are horseshoe-shaped and there are 4 sub- genital pits. Both the gonads and subgenital pits are smaller than in Aurelia flavidula. The gelatinous substance of the bell is bluish white in color. The genital organs of the males are pink, and of the females white in color. The basal bulbs of the marginal tentacles are often pink. This medusa is extremely abundant in Havana Harbor, Cuba, in February. It makes its appearance at the Tortugas in August, but we do not know whether it is found also at Havana at that time or not. While in Havana Harbor in February, 1893, we had the opportunity of ob- serving the curious habits of this medusa. During the morning hours not one was to be seen, but at about four o’clock in the afternoon they began to appear in great numbers, and continued to be seen until long after nightfall. This species is quite distinct from Aurelia marginalis, L. Agassiz (1862; 70 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. p- 86). In Aurelia marginalis the gonads are very large and occupy at least one half of the whole diameter of the disk, so that the distance from the peri- pheral outline of these organs to the margin of the disk is as great, if not greater, than that to the centre of the disk. The mouth-arms, on the contrary, are comparatively small. In Aurelia habanensis the gonads never occupy more than one third of the diameter of the disk, and the mouth-arms are long and slender. CHARYBDBHA, Péron and Lesvevr, 1809. Charybdea aurifera, nov. sp. Figs. 81-83, Plate 25. Specific Characters. — Young medusa. Only one specimen of this medusa was found at the Tortugas, Florida, August 6, 1898. The bell was 2 mm. in height and a little higher than it'was broad. The external surface of the bell was covered irregularly with numerous wart-like clusters of nematocyst cells. The 4 interradial tentacles were evidently very immature, and consisted of small knob-like protuberances from the bell margin. They were hollow and were in communication with the general gastro-vascular cavities of the bell. The 4 sense-organs, or rhopalia, arose from 4 radially situated niches, found upon the sides of the bell at a little distance above the margin. It should be noted, however, that although the rhopalia appear to arise at some distance above the bell margin, they are morphologically homologous with appendages of the bell margin. Each rhopalium arises from a niche in the side of the bell, and consists in a stalk-shaped body, bearing upon its distal end a knob- shaped portion which, in turn, contains the otolith and eye-spots. There are 5 eyes in each rhopalium; one of these is large and median, and the other 4 are smaller and paired (see Figures 82, 83). They are so situated that they may look inward towards the bell cavity. These eyes are ectodermal structures, and possess a lens and a layer of pigment cells. The otolith, on the other hand, is entodermal in origin and consists in a mass of glistening white granules. The velarium is well developed, and is supported by means of 4 partitions, or frenule (f, Figure 81), that suspend it from the sub-umbrella. The proboscis is wide and flask-shaped, and there are 4 quadratic lips. 4 long gastric cirri, one in each interradius, extend downwards into the stomach cavity. A highly refractive band of muscle fibres ? (m s, Figure 81) extend down the middle line of each radius of the bell to the rhopalia. The gelatinous sub- stance of the bell possesses a bluish tinge. The nematocyst cells of the ex- umbrella, and also the proboscis, rhopalia, and tentacles are of a decided amber color. Although careful search was made for them, no velar canals were observed, It is probable that these may develop at a later stage. MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. vb: Charybdea punctata. Tamoya punctata, Fewkes, J. W., 1883, Bull. Mus. Comp. Zool. at Harvard Coll., Vol. XI. p. 84, Figs. 4-6, Pl. I. A single young medusa of this species was found at the Tortugas, Florida, on May 24, 1899. It was very nearly in the same state of development as the meduse described by Fewkes, 1883, from the Bermuda Islands. The adult medusa has not been found. CASSIOPEHA, Péron and Lesveur, 1809. Cassiopea frondosa, Lamarck. Medusa frondosa, Pallas, P. S., 1774, Spicilegia Zodlog., Fasc. X. pp. 29, 30, Pl. 2, Figs. 1-3. Cassiopea frondosa, Lamarck, J. de, 1817, Hist. Nat. Anim. sans Vert., Tom. II. p. 512. Cassiopea pallasii, Péron, F., et Lesueur, C. A., 1809, Tableau des Meduses, p. 357, Nr. 88. Polyclonia frondosa, Agassiz, L., 1860, Cont. Nat. Hist. U. S., Vol. III. Pls. 13, 188, This medusa is very abundant at the Tortugas and along the Florida Reefs early in the spring, but is not seen during the summer months. Ill. SIPHONOPHOR. VELELLA, Bosc, 1802. Velella mutica, Bosc. Medusa velella, Linné, 1767, Systema Nature, Ed. XII. p. 1098. Velella mutica, Bosc, L. A. G., 1802, Hist. Nat. d. Vers., Tom. IL. p. 158. Velella mutica, Agassiz, A., 1883, Mem. Mus. Comp. Zodl. at Harvard Coll., Vol. VIII. No. 2, p. 2, Pls. I.-VI. 91 Figures. Armenista mutica, Haeckel, E., 1888, Siphonophore, Challenger Report, Zodl. Vol. XXVIII. p. 84. This Siphonophore appears occasionally in great numbers at the Tortugas, Florida, especialiy when southerly breezes drive the surface waters of the Gulf Stream upon the Florida Reefs. It is common among the Bahama Islands and along the Cuban coast, and isolated individuals are often carried far to the northward by the Gulf Stream, specimens having been taken in Newport Harbor, Rhode Island. TZ BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. PORPITA, Lamarcr, 1816. Porpita Linnzeana, Lesson. Porpita Linnzana, Lesson, R. P., 1843, Hist. Nat. des Zooph. Acal., p. 588. This Siphonophore is met with occasionally at the Tortugas, Florida. It often occurs in vast swarms, which appear at irregular intervals, all along the coast of the United States from the Tortugas to North Carolina. A single specimen was found by A. Agassiz, in Newport Harbor, Rhode Island, in 1875. RHIZOPHYSA, Prron and Lesvevr, 1809. Rhizophysa Murrayana, Cuun. Rhizophysa filiformis ? Gegenbaur, C., 1854, Zeit. fiir Wissen. Zodl., Bd. V. p. 324, Taf. XVIII. Figs. 5-11. Cannophysa Murrayana, Haeckel, E., 1888, Siphonophore, Challenger Report, Zodl., Vol. XXVIII. p. 324, Pl. XXIV. Figs. 1-9. Cannophysa Eysenhardtii, Mayer, A. G., 1894, Bull. Mus. Comp. Zoél. Harvard Coll., Vol. XXV. p. 239, Pl. III. Figs. 1, 2, 4. Rhizophysa Murrayana, Chun, C , 1897, Siphonophoren der Plankton Expedition, p. 84. This Siphonophore has been found by us among the Bahama Islands and off the Cuban coast, and a single damaged specimen was obtained at the Tortugas, Florida. It has been obtained by Haeckel, 1888, at the Canaries, and by Chun, 1897, in the Tropical Atlantic. The Mediterranean species R. filiformis of Gegenbaur, 1854, is certainly very closely allied, if not identical with the Atlantic form. Rhizophysa Eysenhardtii, Gecrenzaur. Rhizophysa filiformis, Huxley, T. H., 1859, Oceanic Hydrozoa, p. 90, Pl. VIII. Figs. 13-20. Rhizophysa Eysenhardtii, Gegenbaur, C., 1859, Nova Acta Acad. Nat. Curios., Tom. 27, p. 408, Taf. 31, Fig. 46-49. Nectophysa Wyvillei, Haeckel, E., 1888, Siphonophore, Challenger Report, Zodl., Vol. XXVIII. p. 327, Pl. XXIII. Figs. 1-8. This Siphonophore is found occasionally at the Tortugas, Florida. It has been obtained by Haeckel in the Canary Islands, by the Plankton Expedition in the Sargasso Sea, and by Fewkes in the Bermudas. Rhizophysa clavigera, Cuoun. Cannophysa filiformis, Mayer, A. G., 1894, Bull. Mus. Comp. Zool. at Harvard Coll., Vol. XXYV. p. 241, Pl. III. Fig. 3. Rhizophysa clavigera, Chun, C., 1897, Siphonophoren der Plankton Expedition, p. 104. A single specimen of this Siphonophore was found floating within ten metres of the surface in the Gulf Stream between Havana and Key West. MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 73 PHYSALIA, Bosc, 1802. Physalia pelagica, Bosc. Salacia phisalus, Linné, 1756, Systema Nature, p. 158. Holothuria physalis, Linné, 1767, Syst. Nature, Ed. XII. p. 1090. Medusa caravella, Miiller, O. F., 1776, Besch. Berlin Gesell. Natur. Freunde, Bd. IL. p. 290, Taf. IX. Fig. .1. Arethusa crista subrubella venosa, Browne, P., 1789, Nat. Hist. Jamaica, p. 386. Physalia pelagica, Bosc, L. A. G., 1802, Hist. Nat. d. Vers., Tom. II. p. 168. Physalis arethusa, Tilesius, W. G., 1812, Krusenst. Reise, p. 91, Pl. XXIII. Figs. 1-6. . Physalia caravella, Eschscholtz, F., 1829, Syst. der Acalephen, p. 160, Taf. XIV. Physalia aurigera, McCrady, J., 1857, Gymn. Charleston Harbor, p. 74. Physalia Olfersii, Quatrefages, A. de, 1854; Ann. des Sci. Nat., Ser. 4, Zool., Tom. IT. p. 112, Pl. IT. Figs. 1-9; Pl. IV. Figs. 1, 2. Caravella maxima, Haeckel, E., 1888, Siphonophore, Challenger Report, Zodl., Vol. XXVIII. pp. 3138, 338, 352, Pl. XX VI. Fig. 8. Physalia maxima, Goto, S., 1897, Journ. Coll. Sci. Imperial Univ., Tokyo, Japan, Vol. X. Part II. p. 175, Taf. XV. Figs. 1-12. This large Siphonophore is frequently seen throughout the year floating past the Islands of the Tortugas, Florida, and a southerly breeze is almost sure to strand large numbers of them upon the beaches. The animal is found all over the Tropical and Subtropical Atlantic. It is carried by the Gulf Stream to the shore of Europe, and is often found in the Mediterranean near the Straits of Gibraltar. It appears, however, not to be permanently established in the Mediterranean. During the latter part of the summer this Siphonophore is quite common along the southern coast of New England, and individuals have been found as far north as the Bay of Fundy. SPHAITRONECTES, Hoxtey, 1859. Spheronectes gracilis, Harcket. Fig. 89, Plate 27. Monophyes gracilis, Claus, C., 1874, Schrift. Zod]. Inst. Wien, II. Die Gattung Monophyes, p. 29, Taf. IV. Figs. 8-14. Spheronectes inermis, Fewkes, J. W., 1880, Bull. Mus. Comp. Zool. at Harvard Coll., Vol. VI. p. 148, Pl. II. Fig. 6. This Siphonophore is abundant in the Mediterranean and Tropical Atlantic. It was found throughout the winter by Chun in the Canary Islands, and sev- eral specimens were found by us at the Tortugas, Florida, in July, 1898. A 74. BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. single specimen of its sexual generation (Diplophysa inermis) was found by Fewkes (1881; Bull. Mus. Comp. Zool., Vol. VIII. p. 166, Plate VI. Figure 12), in Newport Harbor, Rhode Island. The Atlantic species of Spheronectes is closely allied to, if not identical with, S. Kollikeri of the Tropical Pacific. Chun (1892, Abhandl. Senck. Gesell., Bd. 18, p. 86) says that in 8. Kollikeri the distal portion of the phyllocyst curves downward toward the edge of the swimming-bell, while in S. gracilis it bends upwards. Our observations on §, Kollikeri from the Fiji Islands do not support this view (see Agassiz and Mayer, 1899 Bull. Mus. Comp. Zool. at Harvard Coll., Vol. XXXII. p. 177, Plate 16, Figure 51), for there appears to be much individual variability in respect to the curvature of the phyllocyst in the Pacific species, All of the specimens of S. gracilis observed in the Tortugas, Florida, were colorless, whereas S. Kollikeri is often quite highly colored ; the entoderm of the feeding-polypites being bright yel- low, and the nematocyst batteries of the tentacles orange. { DIPHYES, Cuvier, 1817. UBUDOXIA, Escuscuorrz, 1825. eat i bipartita, Cosra. Eudoxia campanula, Leuckart. Figs. 114, 114", Plate 34. Diphyes bipartita, Costa, O. G., 1840, Genere Diphya, p. 4, Taf. IV. Diphyes acuminata, Leuckart, It., 1853; Die Siphonophoren, p. 61, Poly gastric Taf. III. Figs. 11-19. Generation Diphyes gracilis, Gegenbaur, C., 1854, Zeit. fiir Wissen. Zool., Bd. V. p. 309, Taf. XVI. Figs. 5-7. Diphyes Sieboldii, Kélliker, A., 1853, Die Schwimmpolypen der Messina, p. 36, Taf. XI. Figs. 1-8. This species is very abundant all over the Tropical Atlantic and in the Mediterranean ; and specimens are often found at Newport, Rhode Island, late in the summer. MAYER: MEDUS#Z FROM THE TORTUGAS, FLORIDA. 19 Pe leaaene HAECcKEL, 1888. ERS AVA, Escuscuo.rz, 1829. Diphyopsis campanulifera, Cuun. Figs. 93-95, Plate 28. Diphyes —, Quoy, J. R. C., and Gaimard, P., 1827, Ann. des Sci. p. 137. Diphyes Bory, Quoy, J. R. C., and Gaimard, P., 1833, Voyage de l’ Astrolabe, Tom. IV. Zoéphytes, p. 83, Pl. IV. Figs. 1-6. Diphyopsis campanulifera, Chun, C., 1888, Sitzungsber. Akad. Wissen. Zool., Bd. XLIV. p. 1159. Diphyopsis compressa, Haeckel, E., 1888, Siphonophore., Challen- ger Report, Zool., Vol. XX VIII. p. 153, Plates 33, 34, 18 Figs. Polygastric ( Nat., Tom. 10, Pl. I. Fig. 7. Diphyes campanulifera, Eschscholtz, F., 1829, Syst. der Acalephen, Generation Erseza Lessonii, Cuun. Figs. 96, 97, Plate 28. { Erseea Gaimardi, Eschscholtz, F., 1829, Syst. der Acalephen, p. 128, Taf. XII. Fig. 4. Free Sexual Eudoxia Lessonii, Huxley, T. H., 1859, Oceanic Hydrozoa, p. 57, Generation Pl. III. Fig. 6. Ersea compressa, Haeckel, E., 1888, Siphonophore, Challenger Report, Zool., Vol. XXVIII. p. 123, Pl. XXXIV. Figs. y-18, This Siphonophore is common all over the Tropical and Subtropical Atlantic. A few individuals are drifted into Newport Harbor every summer by the southerly winds, and are probably blown northward from the waters of the Gulf Stream. Diphyopsis picta. Doramasia picta, Chun, C., 1888, Sitzungsber. Akad. Wissen. Polygastric Berlin, Bd. XLIV. p. 1154. Generation Doramasia picta, Chun, C., 1892, Abhandl. Senckenberg Gesell., it Bd. XVIII. p. 91, Taf. VIII. Fig. 3; Taf. IX. Figs. 5-9. Ersza picta, Cuun. Fig. 118, Plate 34. Free Sexual Ersza picta, Chun, C., 1888, Sitzungsber. Akad. Wissen. Berlin, Generation ; Bd. XLIV. p. 1154. This form has been found by Chun in the Canary Islands, and it is also common at the Tortugas, Florida, We present a figure of the free sexual generation. VOL. XXXVII.— NO. 2. 6 76 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. ; Diphyopsis hispaniana,! nov. sp. Erseza hispaniana, nov. sp. Figs. 98-99, Plate 29. Polygastric generation = Diphyopsis hispaniana. Monogastric, sexual generation = Ersewa hispaniana. Specific Characters. — Diphyopsis hispaniana. Figures 98,99. The animal is 12 mm.in length. The cavity of the anterior swimming-bell is very volumi- nous, so that the bell walls are remarkably thin. They are, however, quite rigid, so that swimming is accomplished almost entirely by the movements of the powerful velum. There is a well-developed hydreecium upon the ventral side of the anterior swimming-bell, and a long spindle-shaped phyllocyst arises from its inner apex, and extends upwards along the side of the bell cavity. The siphosome arises from the inner apex of the hydrecium, immediately under the point of origin of the phyllocyst. The first appendage of the siphosome is the large posterior swimming-bell that is almost as large as the anterior. It is provided with 4 radial canals, and a circular vessel, and these are placed in connection with the gastro-vascular space of the siphosome by means of a long slender duct. The posterior swimming-bell possesses a well-developed velum, the contractions of which acting simultaneously with those of the velum of the anterior swimming-bell, cause the animal to dart through the water at a very rapid rate. The posterior swimming-bell is provided with two large lateral wings having serrated edges. The siphosome extends downward through the groove between these wings. The order of development of the various organs upon the siphosome is as follows : — First the feeding-polypites, then the ten- tacles, then the gonads and swimming-bells, and lastly the covering scales. The feeding-polypites are spindle-shaped, and quite contractile. The outer surface of their proximal portions displays a number of wart-like swellings. The entodermal cells of these swellings are of a decided ochre-yellow color, and it seems not improbable that their function may be similar to that of the “liver cells’ of the feeding-polypites of Agalma. The tentacles arise from the sides of the feeding-polypites very near their point of origin from the siphosome. They give rise to a number of lateral branches that are studded with sharply projecting nematocyst cells. (See Figure 99.) These lateral branches term- inate in swollen nematocyst batteries. The covering scales are spathiform and possess a deep ventral groove. A single gonad and a swimming-bell bud out side by side, very close together, from the base of each feeding-polypite. When sufficiently developed, each unit, consisting in a feeding-polypite, tentacle, gonophore, swimming-bell and covering scale, is set free from the siphosome of Diphyopsis hispaniana, and becomes the free-swimming, mono- gastric, sexual generation known as Ersea hispaniana. 1 Called “hispaniana ” on account of its red and yellow coloration. The ento- derm of the feeding-polypites being ochre-yellow, and the tentacular nematocyst- batteries port-wine-red. MAYER: MEDUS# FROM THE TORTUGAS, FLORIDA. 77 Erszea hispaniana, nov. sp. Fig. 100, Plate 29. Specific Characters. — Erszea hispaniana. The mature animal is 7 mm. in length. The covering scale is hood-shaped without a sharp apex. Its lower portion is sharp-edged, and overlaps the large swimming-bell. The phyllocyst is short and blunt, and contains a highly refractive “oil” globule. The large swimming-bell is provided with 4 longitudinal, serrated ridges that give ita rectangular appearance in cross-section. There are 4 radial tubes, a circular vessel, and a well-developed velum. Two or more gonophores are seen bud- ding out from the side of the feeding-polypite near its base. These gonophores (g, Figure 100) are medusiform and are provided with 4 radial tubes, a circular vessel, anda velum. The genital products are found within the manubrium. In Figure 100 a single large, oval egg is seen occupying this position. Both Diphyopsis hispaniana and Ersea hispaniana were common at the Tortugas, Florida, in July, 1898, but were not seen during the summer of 1899. es Quoy and Garmarp, 1827. AGLAISMA, Escuscuotrz, 1829. Abyla pentagona, EscuscHotrrz. Figs. 101, 1012-103, Plate 30. Abyla pentagona, Eschscholtz, F., 1829, Syst. der Acalephen, p. 132. Calpe pentagona, Quoy, J. R. C., and Gaimard, P., 1827, Ann. der Sci. Nat., Tom. X. p. 11, Pl. 2 A. Figs. 1-7. Polygastric Abyla trigona, Vogt, C., 1854, Mém. de l'Institut Nat. Génevois, Generation Tom. I. p. 121, Pl. XX. Figs. 4-7; Pl. XXI. Figs. 3-6, 10-18. Calpe Gegenbauri, Haeckel, E., 1888, Siphonophore, Challenger Report, Zoél., Vol. XXVIIL. p. 164, Pl. XX XIX. Figs. 1-12. Abylopsis pentagona, Chun, C., 1897, Siphonophoren der Plank- ton Expedition, Bd. II. K.b. p. 30. Aglaisma cuboides, Cuun. Fig. 104, Plate 30. Eudoxia cuboides, Leuckart, R., 1853, Siphonophoren, p. 59, Taf. III. Figs. 7, 8, 10. Einzelthiere der Abyla pentagona, Gegenbanr, C., 1854, Zeit. fiir Free Sexual Wissen. Zool., Bd. V. p. 295, Taf. XVI. Figs. 1, 2. Generation Aglaisma Gegenbauri, Haeckel, E., 1888, Siphonophore, Chal- lenger Report, Zo6l., Vol. XXVIII. p. 119, Plate XL. Figs. 13-20. Aglaisma cuboides, Chun, C., 1897, Siphonophoren der Plankton Expedition, p. 30. 78 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. This form is quite common at the Tortugas, Florida, as indeed it is all over the Tropical and Subtropical Atlantic. It is found in Charleston Harbor, South Carolina, but has not yet been taken north of the Carolina coast. Abyla quincunx, Cuun. Figs. 115-117, Plate 34. Abyla pentagona, Huxley, T. H., 1859, Oceanic Hydrozoa, p. 40, Polygastric Pl. IL. Figs. 2-2e. Generation 1 Abylopsis quincunx, Chun, C., 1888, Sitzungsber. Akad. Wissen. {| Berlin, Bd. XLIV. p. 1160. Aglaisma quincunx. f Aglaismoides Eschscholtzii, Chun, C., 1888, Sitzungsber. Akad. Free Sexual Wissen. Berlin, Bd. XLIV. p. 1160. Generation Aglaismoides quincunx, Chun, C., 1897, Siphonophoren der Plank- ton Expedition, p. 29. Chun found this Siphonophore in the Canary Islands, and it was taken by the Plankton Expedition in the Gulf Stream and Sargasso Sea. Huxley, 1859, found it in the tropical regions of the Atlantic, Pacific, and Indian Oceans. It has been taken by Agassiz and Mayer (1899; Bull. Mus, Com. Zool., Vol. XXXII. p. 180) in the Fiji Islands. These South Pacific speci- mens are, however, slightly different from those of the Atlantic in that their tentacular nematocyst-batteries are usually colorless instead of more or less orange, as in the Atlantic form. Chunia capillaria, nov. gen. et sp. Figure 90, Plate 27. Generic Characters. —Chunia, novum genus. This genus belongs to the family Diphyide, Eschscholtz, and to the subfamily Abylinew, L. Agassiz. It possesses a pentagonal, prismatic, anterior swimming-bell and a larger five- sided, posterior swimming-bell. The siphosome bears a long, slightly curved, sharp-pointed, hair-like bristle. The covering scales, or bracts, are leaf-shaped. The monogastric sexual generation is unknown. Specific Characters. — The animal is about 10 mm. in length. The anterior swimming-bell is prismatic, and possesses one oblique, five-sided face and 5 lateral faces. Four of these are plane, but the fifth is sharply concave. The cavity of the swimming-bell opens upon this concave face. The bell cavity is long and spindle-shaped, and is provided with 4 radial tubes and a velum, There is a large spherical phyllocyst that gives rise to an apical caecum con- taining a highly refractive ‘‘oil” globule. The hydroecium of the anterior MAYER: MEDUSA FROM THE TORTUGAS, FLORIDA. 79 swimming-bell is long and tube-like, and its axis is parallel to that of the bell cavity. Indeed, it is quite similar in form to the hydrecium of Abyla quincunx (see Figure 115, Plate 34). The first appendage of the siphosome is the posterior swimming-bell. This is somewhat larger than the anterior and is five-sided, the sides being bounded by prominent, angular, serrated ridges. There is a single median dorsal ridge and two pairs of lateral ridges. The ventral-most pair of lateral ridges are wing-like, and enclose a trough-like groove through which thesiphosome extends. The bell cavity of the posterior swimming-bell is spindle-shaped and is provided with 4 radial tubes and a circular vessel. These are placed in communication with the general gastro- vascular cavity of the siphosome by means of a long slender duct. The posterior swimming-bell is furnished with a powerful velum, by the con- tractions of which the animal is enabled to shoot through the water. The siphosome is not very long and rarely extends beyond the posterior extremity of the trough-like groove in which it lies. The first organs to be developed upon it are the feeding-polypites; the tentacles soon arise as buds from the sides of the polypites, each feeding-polypite being provided with a single tentacle. The tentacles give rise to lateral branches each one of which termi- nates in a swollen cylinder-shaped nematocyst battery. A long, slightly curved, bristle-like spine arises from the siphosome at a short distance below its origin, and extends outward to a considerable distance beyond the distal end of the posterior swimming-bell. It seems probable that this structure may be morphologically equivalent to a bract, that has become thus modified for defensive purposes. Other covering scales or bracts were observed upon the siphosome, but these were leaf-shaped (see cs, Figure 90, Plate 27). No gonophores or sexual organs were observed, and the sexual generation is unknown. This rare form is found among the Bahama Islands during the winter months. The specimen from which we have obtained our figure was captured in Nassau Harbor, New Providence Island. We also obtained a specimen at Watlings Island (San Salvador) on January 15, 1893. AGALMA, Escuscuotrz, 1825. Agalma Pourtalesii, Acassiz and Mayer. Figures 106-113, Plates 31-33. Agalma Pourtalesii, Agassiz, A., and Mayer, A. G., 1899; Acalephs from the Fiji Islands, Bull. Mus. Comp. Zool. at Harvard Coll., Vol. XXXII. p. 180. Spectfie Characters. — The entire animal (Plate 32) is about 25 mm. in length. The feeding-polypites, dactylozoids, tentacles, and gonostyles, all arise from the ventral side of the siphosome. The float, or pneumatophore, is of small size, and its apical pore is surrounded by radially arranged streaks of dark 80 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. red pigment. Theswimming-bells (Figures 109, 110, Plate 33) are dovetailed alternately, one above another, so that their velar openings are situated on two diametrically opposite sides of the nectosome (see Figure 108). The siphosome is densely covered upon all sides with thick prismatic bracts, or covering scales. One of these bracts detached from the animal is shown in Figure 112. The angular edges of the older bracts are usually smooth, but in the younger ones they frequently display a row of regularly arranged nema- tocyst-bearing papille (see Figure 113). 2 iy oie wl" : 7 : . ‘ -~ a *® r eoubels Gaguiie’l — s574M i . ; - . A : ' - ' ‘ s ri. . tel Ps, ee ir a a =~ ] 3 = FS 4 ed 7 A "; . e y iP 44aelra Sos es- , ul (Th ‘ : ' ri tig Thee Yi = ' bod cleats as fas c= ’ ‘ ‘ | ’ F ote 4 “@ PLATE 1. Matt Pandea violacea, Agassiz and Mayer. Stomotoca australis, nov. sp. i 1 ie Wt adal ~~ - a { é ss = : Zi i; ee ’ “uoisgr [astayl I ‘PRWIV ri o = i rar i , a, pe 7 "y ny ill ; : 5 ( eee weqetesT > onal : | Ly TTA Phé os ron wbirust stsonomid = b aid gown ‘ bs it 4 iY Ms 5 BN + if \ 9 Mayer. — Tortugas Meduse, * f t Ay ar Ai * - PLATE 2. — Fig. 3. Dissonema turrida, nov. sp. Adult medusa. Fig. 4. vee & Young medusa. = : “nsOR"yy TAStayy PRWOYV S ° iiviel aependT — davar P 7 ' 6 STAI ge Vien ecitoenerdt BH itl a pot of yi Mo nobel. 40 fo ery Oo alidqneg sehinaigauinol) 2 ait eebaatod yitievd-inyomanior bia yamaut-raiowe pe : wrodd wlilqouy «shinies q 4 ‘ ‘ . = : 7 f = Oe ie Mayer. — Tortugas Meduse. PLATE 3. 5. Bougainvillia frondosa, nov. sp. 6. Gonionemoides geophila, nov. gen. et sp. Portion of bell margin show- ing sucker-bearing and nematocyst-bearing tentacles. 7. Gonionemoides geophila, otocyst. ~\ A \ yy 5 B Meisel hth Beste. f ® : iret f i |) ’ . a gle. ‘ ; Da yy Bae ' ' p avs “ Var ' = + whoa \ t f ! ) p » A i | 4 io 7 m4 aA 4 ba oe : y) D «7 re ; ed Ble ‘ Ty et } Vie sy ' 7 t 7 4 e F ° f ‘ t : 7 5 i its 1% ‘ ‘ i id : iy i} ‘ ’ : } : iw }s ‘ i , iy ( j Fi ‘ } 7 7 - a ——— - a 7 _ , 7 7 ’ P ofS yf, Mayer. — Tortugas Meduse. PLATE 4. Fig. 8. Gonionemoides geophila, nov. gen. et sp. Side view of adult medusa. Fig. 9. 5 £ Medusa with oral surface upward, and holding fast to the bottom of the aquarium by means of the adhesive suckers upon its tentacles. 9 oe a imaghal! qe ect su ut : dab'l Reilen gover be te > yeni ae Nye ,bhelysgy Wt hjoare ti, iid Ae 70m Bie Sel Coorleridaers. SD lateincy asl Vs, soldancidt af) Ww antinigys Prey ipst Hod Yes this awed sida goed ee Le . : on 7 es Mayer. — Tortugas Meduse. PLATE 5. Fig. 10. Gonionemoides geophila, nov. sp. Young medusa. Fig. 11. Gonionemoides geophila, nov. sp. Distal end of tentacle showing the beginning of the formation of the suctorial disk. Fig. 12. Halicalyx tenuis, Fewkes. Portion of bell margin. a ia i B Meisel lith, Boston. met ah gi Aiareae 7 wayne 7 ' ra he. feaptun YW Weil abt oS ed # TIRE: PLATE 6. ‘ Fig. 13. Halicalyx tenuis, Fewkes. Side view of mature mei lus ‘ ‘ ‘9 HTLV Ig “SVSNdaAW SVONIYNOT ) . SBS 5s B Meisel lith, Boston, AGM. del, a te cise harh aaiggnrv'l — cmv eM Ps 7 \ ; we a >. Fig. 15. “ee PLATE 7. Puate 7. B Meisel lith. Boston. % vy mug ~~ sea Le oe ) 7 6 ATAL) : AY 4 Te Roy Sor baile einweortiaT OL yt : 7 MW eritibaney Sil —wivluninsy stounmaytl 911 pl _ a J is i : m) , q i , b - ‘ é r a e , ° a - Mayer. — Tortugas Meduse. PLATE 8. Puate. 8. B Meisel lith Boston. i) ny ‘ _ " Ki Y ®@aTaAs F a8 ile tiga Ned Jo nodW) we, Vie 0TH AQ nhyrnet Uni Ba woiste'l i Oithay as DEMO Melt hoy os rs 7 Mayer. — Tortugas Medusez. . PLATE 9. Fig. 18. Oceania magnifica, nov. sp. Fig. 182. $ Portion of bell margin, showing otocysts. Fig. 19. Oceania globosa, nov. sp. Fig. 198. ee “Portion of bell margin. - rs B Meisel lith,Bostea. OL STAT : » ipl. you anonitaiog riecstou > ‘ae ig7nct) - al MAH e neiligl = alablinartg sitvod'l , migra iat ton rir 4 } 2 ng 3 : i (ee oi «ihy xof { 1X td Pagel gil Mayer. — Tortugas Meduse. PLATE 10. Fig. 20. Oceania gelatinosa, nov. sp. Fig. 202. s ss Otocyst. Fig. 21. Phortis pyramidalis = Eutima pyramidalis, L. Agassiz. Fig. 212. oS ss Portion of bell margin. a) it fe Ei < = Ay = a ae “<< mA « (atte eeguiwT — ose. F 1 a ¢ “i cs aTasT ae ak .epirta giro = aalgiies edqiamitie ok hie: ppipinn fled to niizw") ei ni ap wien? =«6cisteyA a) aetiys sleaiopyS.. CE it ‘ : or ategon) ae aya are 7 ’’ : . . ; s! “il cat - a : Mayer. — Tortugas Meduse. PLATE 11. Fig. 22, Eutimalphes cerulea = Eirene ccerulea, L. Agassiz. Fig. 228. . a Portion of bell margin. Fig. 23. Zygodactyla cyanea, L. Agassiz. Young medusa. Fig. 232. iy s Otocysts. = fx Es 7 < Boil Ay B.Metsel lith Bosten, ' a ae ' ao 7 i. - i 7 : e* i f ‘ : 4 n x . teed eegisee Tp arta ; 4 . : 7 . , p = 7 . G | * 5 * Bi Waal : yebje HA PO. ROA FOU wie niga; Mad vinby oad ee F Rapbedi yo weivimi? = ™ ro et MSYI tO.wes? was , : Nae hgh , ne 4 7 ” is —————7~ ‘4 i) Mayer. — Tortugas Medusz. Fig. 24. Pseudoclytia pentata, Nov. gen. et sp. Side view. — Fig. 25. A Fig. 26. s PLATE 12. “Oral view of medusa. “Side view of otocyst. g a 3 a = mS ; ict la i L ot! £ ii " ii { pe ay Ws Or ae Ak A 2 | pl bo kegel » wars tl j ; : fa | u PATS Bawale chet: ate ron em Sng taoonlll 72 yt Mende to yoty fax . i att * Neieoos re Mayer. — Tortugas Meduse. PLATE 13. Fig. 27. Rhacostoma dispar, nov. sp. Side view of medusa. Fig. 28. s “Oral view of medusa. Fig. 29. af “ Otocyst. hi aa 13: B. Meisel lith, Boston sentbentt cme iret f ) $1 STATS 4 5 aly 4618. dodooal! nnueciesih allouigds fnatbodlgaet i ARQ ABV! {rabyael, ae » a, + os 7 * _ - . = een Mayer. — Tortugas Meduse. Fig. 30. Fig. 31. Fig. 32. PLATE 14. /EHginella dissonema, Haeckel. Side view of medusa. se of Longitudinal section of tentacle. s ef Marginal sense-organ, PLate 14. ert B Meisel lith, Boston. 4 Li « - “4 0 i ‘ i F <3 v) 7 = a uP) Marte eeqerrsT ANT 4 ! es TASS hi file renbelf Planes 7 wenaye 3 | Uromtroy¢S J yi 4 iigieant Sho to noibrol Wi bf. gid ae 1 7" ’ ; - an yy ete ber te te. en VO Gieinog Hil ebro tiene ug ig yuary Yo WHY 7 Mayer. — Tortugas Meduse. PLATE 15. Fig. 33. Zygodactyla cyanea, L. Agassiz. Medusa with 16 radial tubes. Fig. 34. 5 “Portion of bell margin. Fig. 85. Pseudoclytia pentata, nov. gen. et sp., red variety. Fig. 35*. me £ View of ovary. Ww Loo | t E < 4 A, TUC SAS MEDUSAE. B Meisel lith Boston. eth eran eel of TTA if Pee” axoorll wlikety oiquianusie AS ont Aaa Dare yauoy tS ait (¢ Vol evaedalat wrote 4 ee Th | Qo Vor edroqus mall wait Mayer. — Tortugas Meduse. PLATE 16. Fig. 36. Steenstrupia gracilis, Brooks. Mature medusa @. Fig. 37. ot Young medusa. Fig. 38. Ectopleura minerva, nov. sp. Fig. 89. Tiara superba, nov. sy “ec al - PLaTeE. 16. B Metsel lith Boston. papel’ aagirivet ~ cara? ~ Py oe We a a qe vor) Mmiitoitos ) pri pater santas) fe ape Vor wilieert anvteghT th yw Li} y by 7 > % : melee. fp » r Se bol . ban va 4 Mayer. — Tortugas Medusz. PLATE 17. Fig. 40. Gemmaria dichotoma, nov. sp. Fig. 41. Dipurena fragilis, nov. sp. o * w 7. PLATE 1 eas 2S. Res! B Meisel lith,Bosten. * an! ¥ - A A .7 a4 = nA . ea , c. sf , BL ATAST je wor ch mn neiherrqundey mate. 99 -yor cousaldonsed moe Pihird reap Pa #1 wa lnwdA qe vor om aorta) nat Yo melting lanl oy dodoonl) adiwwasentiol simvbowns Mayer. — Tortugas Meduse. Fig. Fig. Fig. Fig. Fig. Fig. 42. 43, 44. 45. 46. 47. PLATE 18. Dysmorphosa minuta, nov. sp. Netocertoides brachiatum, noy. gen. et sp. Side view. ss cs Aboral view. Dipurena picta, nov. sp. # “Terminal portion of tentacle. Staurodiscus tetrastaurus, Haeckel. Young medusa. Puate 18. ORTUGAS MEDUSAE. & oS > Me, se B Meisel ith, Boston. i MAYER, — Tortugas Meduse, PLATE 19. Fig. 48. Staurodiscus tetrastaurus, Haeckel. Oral view of mature medusa. Fig. 49. 7: 2 Side view of mature medusa. PLATE. 19. B Meisel ith, Basten. Mayer. — Tortugas Meduse. Fig. Fig. Fig. Fig. Fig. Fig. PLATE 20. Laodicea neptuna, nov. sp. View of tentacle and ocellus. ‘ Side view of mature medusa. Oral view of mature medusa. Oceania discoida, nov. sp. Side view of mature medusa. ss oh Side view of young medusa. Oral view of bell margin of a young medusa. “ “c “ec “ PLATE. 20. GAS MEDUSAE. “B Meisel lith Basten MAYER. — Tortugas Meduse. PLATE 21. Fig. 56. Oceania McCradyi, Brooks. Side view of mature medusa. Fig. 57. Oceania McCradyi, Brooks. View of one of the hydroid blastostyles that are produced upon the gonads. Fig. 58. Oceania McCradyi, Brooks. Young blastostyle. Fig. 59. 2 2 e Oral view of mature medusa. n PLATE A i AS MEDUSAE. B Meisel lith Bostes. ‘ a i PY : — ° ° . * . : y ‘ . ‘ a *. t a i 1 ‘ ot 7 . , ) fe ‘ = . nt - . - r = . ‘ ' . ' . Sis Mayer. — Tortugas Medusz. Fig. 60. Fig. 61. Fig. 62. Fig. 63. Fig. 64. Fig. 65. Fig. 66. PLATE 22. Tiprepais punctata, nov. sp. Young medusa. ‘¢ Medusa older than Fig. 60. Otocyst of young medusa. Otocyst of medusa drawn in Fig. 61. Dysmorphosa dubia, nov. sp. Side view. de Oral region of the proboscis. Tentacle and ocellus. «“ “ 74 o - - a LJ RTUGAS MEDUSAE. PhaTE 22. B. Meisel lith Boston. ee a y Uf , iy Bhicia fo pon Mayer. — Tortugas Meduse. Fig. 67. Fig. 68. Fig. 69. Fig. 70. Fig. 71. Fig. 72. PLATE 23. Neaehiee punctata, Kolliker. Oral view of young ephyra. i Oral view of sense-organ of young ephyra. Eutimium serpentinum, nov. sp. Mature medusa. e ae View of bell margin, and tentacle. 4 Gastric portion of the proboscis. ee « Proximal portion of the peduncle. PLATE 23. a - , UGAS MEDUSAE. B Meisel lith, Sostes. Mayer. — Tortugas Medus. PLATE 24. Fig. 78. Aurelia habanensis, nov. sp. View of the edge of the mouth-arms, or palps, showing the wart-like clusters of nematocyst cells. Fig. 74. Aurelia habanensis, nov. sp. One of the wart-like clusters of nematocyst cells from the edge of the palps. Fig. 75. Glossocodon tenuirostris, Fewkes. Mature medusa Fig. 76. < < View of mouth. Fig. 77. s ae View of ovary. Fig. 78. xs L View of tentacle and otocyst. PLATE. 24 B Metsel lith,Bosten, — Uy - = = A) : ss, m4 aut | Ezy Phas f ike errs Mayer. — Tortugas Medusm. PLATE 25. Fig. 79. Aglaura hemistoma, Péron and Lesueur. Mature medusa. Fig. 80. a = Otocyst. Fig. 81. Charybdea aurifera, nov. sp. (/), frenula; (ms) muscle strands ? Fig. 82. 4 3 Side view of rhopalium. Fig. 83. i ce View of rhopalium from without the bell. | Fig. 84. Zygodactyla cubana, nov. sp. Fig. 85. o View of bell margin. TORTUGAS MEDUSAE. AGM. det a wrtnt to mati len iit Mayer. — Tortugas Medusz. PLATE 26. Fig. 86. Aurelia habanensis, nov. sp. Oral view of mature Fig. 87. Nausithoé punctata, Kolliker. Oral view of mature m Fig. 88. © s Otocyst and ocellus. s MEDUSAE. B. Meisel lith Sesten. eu Oey Sar eer 16 Mayer. — Tortugas Meduse. PLATE 27. Spheronectes gracilis, Haeckel. (c) connecting canal, (f/) phyllocyst, ( Pp. feeding-polypite, (t) tentacle. Chunia capillaria, nov. gen. et sp. Bolina vitrea, L. Agassiz. Mature animal. : Bolina vitrea, L. Agassiz. View of apical sense-organ. (c) gastric cilia (f) funnel ; (mu) “muscle” fibres. a - | ee ~ fot) el Be ) immature swimming-bell (shown mature in Fig. 96, b). Fig. 95. Diphyopsis campanulifera. Tentacular nematocyst battery. Fig. 96. Ersaa Lessonii, Chun. The monogastric, sexual generation of D. cam- panulifera. (b) swimming-bell; (cs) apical bract, or covering scale ; (g) gonophore. Fig. 97. Ersea Lessonii. Young Q gonophore containing two ova. 7 a PLATE 28. RATUGAS MEDUSAE. — B Meisel ith Basten, - 4 sal = ee ee Mayer. — Tortugas Medusa, if y PLATE 29. Fig. 98. Diphyopsis hispaniana, nov. sp. Bigs 99: 4 ou Portion of the siphosome. Fig. 100. Erswa hispaniana, nov. sp. The monogastric, sexual Diphyopsis hispaniana. (g) gonophore. 7 — >t ae a ne oe a GAS MEDUSAE. PLATE 29. ———a =F “er Vor LN 2 OL B. Meisel {ith Boston IP EE Oe FN IES RE ES TNT a ——E 6A EE LALIT OCLC LY LEELA EOL OLE OL AT LLL EL EE AD IE Mayer. — Tortugas Meduse. PLATE 30. Fig. 101. Abyla pentagona, Eschscholtz. Bigs 10i25) = a Usual form of the phyllocyst. Fig. 102. & es Enlarged view of the siphosome. (as) small “re serve”? swimming-bell ; (cs) bract, or covering scale; (d) duct of th large posterior swimming-bell; (gs) medusiform gonophore. Fig. 103. Abyla pentagona. Tentacular nematocyst battery. Fig. 104. Aglaisma cuboides, Chun. The monogastric, sexual generation of Ab; pentagona. (cs) bract, or covering scale ; (gs) medusiform gonophore that functions also as a swimming-bell. pe PLatTeE 30. a whe ae B. Meisel tith, Bosten. MAyeEr. — Tortugas Medus. PLATE 31. Fig. 105. Ocyroé crystallina, Rang. Figure 14 times the natural size. (i like protuberances upon the surface of the Ctenophore ; (au) Fig. 106. Agalma Pourtalesii, Agassiz and Mayer. Female gonads. Fig. 107. ‘ ss Male gonad. an agave DEY PS \\ \\ we \\i/ S.- \¥ ~ ii} rit HH | \\} y \\ \ Pee Mig, , We AGM. del B Meisel lith, Boston, ja ee ee ee ER ———— SS UCU UC TT Te EW MS a WE Ge Mayer. — Tortugas Medusx. PLATE 32. Fig. 108. Agalma Pourtalesii, Agassiz and Mayer. Side view of the en From a specimen obtained at the Tortugas, Florida. B. Meisel lith, Boston, \ SS ) YQ Ng JGAS MEDUSAE. Pee he ee 27 n V uy TZ MayER. — Tortugas Meduse. Fig. 109. Fig. 110. Fig. 111. Fig. 112. Fig. 113. PLATE 33. Agalma Pourtalesii, Agassiz and Mayer. Swimming-bell seen from — above. Agalma Pourtalesii. Swimming-bell seen from the side. ie o Terminal portion of tentacle. : View of a detached bract. st “ Nematocyst-bearing papilla sometimes seen on young and immature bracts. oT ee aad : — “ AS MEDUSAE Puate. 33. B Meisel lith, Boston. “ B18) Maysgr. — Tortugas Medusz. PLATE 34. Fig. 114. Diphyes bipartita, Costa. Fig. 114°. a One of the units of the siphosome. (g) medusiform gonad, (k) siphosome, (p) feeding-polypite, (t) tentacle, (cs) covering scale or bract. Fig. 115. Abyla SS = Abylopsis quincunx, Chun. Fig. 116. aS Tentacular ee battery. Fig. 117. a ig Small “ reserve ” swimming-bell. Fig. 118. Ersea picta, Chun. B Meisel lith Boston, AGM. del. 7s RA aly Oi Yu 4 au bie teat Tit ati easy df » 4 7 7 : ) Mayer. — Tortugas Meduse. 4 ‘“ PLATE 35. Fig. 119. Agalma virida, nov. sp. Immature individual. aa Big: 120 VS “Terminal knob of the tentacles. Fig. 121. “ “_ Polygonal pigment cells of the float. | se v ee ee roi eo \ Pees slot 5 Meisel lith Boston. « .) -— is Mayer. — Tortugas Meduse. PLATE 36. Fig. 122. Cytaeis gracilis, nov. sp. Mature medusa. Fig. 123. ee cs Young medusa. Fig. 124. fe . Very young medusa. : ; _ B.Meisel lith,Bosten. 36. * MAYER. = Tortugas Meduse. PLATE 37. Fig. 125. Ectopleura minerva, nov. sp. Fig. 126. Dinema jeffersoni, nov. sp. PLATE 37. 7] 7 B Meisel ith. Boston.” Mayer. — Tortugas Medusz. | PLATE 38. 1 Fig. 127. Lizzia elegans, nov. sp. Fig. 128. Eucheilota ventricularis, MeCrady. PLATE 38. ORTUGAS MEDUSAE. B Meisel lith besten. iY Mayer. — Tortugas Meduse. PLATE 39. Figs. 129, 180. Multioralis ovalis, nov. gen. et sp. Fig. 181. Pseudoclytia pentata, nov. gen. et sp. Side view of mature Fig. 152. ss a Tentacle-bulb showing green entoderr PLATE 39. , RTUGAS MEDUSAE. id ‘ G 7 “ +4 aI a B. Meisel Iith Basten. _ , as a : : PO eee ae ee Oe ee eee ’ i Mayer. — Tortugas Meduse. PLATE 40. Fig. 183. Phortis lactea, nov. sp. Fig. 134. Eucheilota paradoxica, nov. sp. Fig. 135. ee if A young medusa bud. Fig. 136. Ke ce A medusa bud about to be set fre parent. —_ eye 7a a Te ge ee I ee ee > ey r yrs eR ii ee = a san * i m_ Pare 40. 2 . : 4 B.Meisel lith. Sosten 4 eee ' re) ; “—ioar S LR Mayer. — Tortugas Meduse. PLATE 41. Fig. 137. Zanclea gemmosa, McCrady. Young medusa of Gemmaria g Fig. 1388. Gemmaria gemmosa, the hydreid stock of Zanclea gemmosa, Fig. 139. Epenthesis folleata, McCrady. Adult medusa. PLate 41, LEE SF i 4 a A B Meisel lith. Boston, : a hoe 7 . , . ts . ’ : | +: : . ‘ s Bi x n = - +. ea =} = 4 < “ yr « 2 & : : - : m, ms ees - ~~ £ 4 oft a . ss . ~— eee , : — = . 4 “ ‘ . < bay “ - = | Ya . , = 4 - aa x Mayer. — Tortugas Medusa. Fig. 140. Fig. 141. Fig. 142. Fig. 143. PLATE 42. Eucopium parvigastrum, nov. sp. Adult medusa. Niobia dendrotentacula, nov. gen. et sp. Side view of an adult medusa. Niobia dendrotentacula. A young medusa recently separated from the adult individual. Niobia dendrotentacula. The proboscis and ova of a mature medusa after the cessation of the medusa-forming process. PLrate 42. UGAS MEDUSAE. B. Meisel lith, Boston, OL 1. Ue ee) ee oe ele = . Neue Beitrage zur Kenntniss der Zelle. II. Theil. Arch. f. mik. Anat., Bd. 37, pp. 685-751, Taf. 38-40. Flemming, W. 93. Morphologie der Zelle und ihrer Teilungserscheinungen. Anat. Hefte, Abth. 2, Ergebnisse d. Anat. u. Entwick., Bd. 3, pp. 65-84. 158 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Flemming, W. 95. Ueber die Structur der Spinalganglienzellen. Verhandl.d. Anat. Gesell., Versamml. 9 (Anat. Anz., Bd. 10, Erganzungsheft), pp. 19-25. Flemming, W. *95*. Ueber den Bau der Spinalganglienzellen bei Saugethieren, und Be- merkungen uber den der centralen Zellen. Arch. f. mik. Anat., Bd. 46, pp. 379-394, Taf. 19. j Guignard, L. 91. Sur lexistence des “ Sphéres attractives” dans les cellules végétales. Comp. Rend. Acad. Sci., Paris, Tome 112, pp. 539-542. Guignard, L. *97. Les centrosomes chez les Végétaux. Comp. Rend. Acad. Sci., Paris, Tome 125, pp. 1148-1153. Gulland, G. L. 96. On the Granular Leucocytes. Jour. Physiol., Vol. 19, pp. 385-417, Pl. 5-6. Hacker, V. 94. Ueber den heutigen Stand der Centrosomafrage. Verhandl. d. deutsch. zool. Gesell. zu Miinchen, 1894, pp. 11-32, 11 Textfig. Hamaker, J. I. 98. The Nervous System of Nereis virens Sars. Bull. Mus. Comp. Zodl., Harvard Coll., Vol. 32, pp. 89-123, 5 pl. Hansemann, D. 91. (Diskussion.) Verhandl. d. Anat. Gesell., Versamml. 5 (Anat. Anz., Bd. 6, Erganzungsheft), pp. 143-144, 2 Textfig. Hansemann, D. °93. Ueber Centrosomen und Attraktionsspharen in ruhenden Zellen. Anat. Anz., Jahrg. 8, pp. 57-59. Heidenhain, M. °91. Ueber die Centralkérperchen und Attractionsspharen der Zellen. Anat. Anz., Jahrg. 6, pp. 421-427. Heidenhain, M. . °92. Ueber Kern und Protoplasma. Festschr. fiir Kolliker, 1892, pp. 111- 166, Taf. 9-11. Leipzig. Heidenhain, M. °94. Neue Untersuchungen iiber die Centralkorper und ihre Beziehungen zum Kern- und Zellprotoplasma. Arch. f. mik. Anat., Bd. 43, pp. 473-758, Taf. 25-31. Heidenhain, M. 97, Ueber die Mikrocentren mehrkerniger Riesenzellen, sowie tiber die Centralkérperfrage im Allgemeinen. Morph. Arbeit., Bd. 7, pp. 225- 280, 20 Textfig. RAND: NERVOUS SYSTEM OF LUMBRICID. 159 Heidenhain, M., und Cohn, T. 97. Ueber die Mikrocentren in den Geweben des Vogelembryos, insbeson- dere tiber die Cylinderzellen und ihre Verhaltniss zum Spannungsgesetz. Morph. Arbeit., Bd. 7, pp. 200-224, 4 Textfig. Henneguy, L. F. 791. Nouvelles recherches sur la division cellulaire indirecte. Jour. de PAnat. et de la Physiol., Année 27, pp. 397-423, Pl. 19. Henneguy, L. F. 96. Lecons sur lacellule. xx+541 pp., 362 fig. Paris. Hermann, F. *91. Beitrag zur Lehre von der Entstehung der karyokinetischen Spindel. Arch. f. mik. Anat., Bd. 37, pp. 569-586, Taf. 31, 1 Textfig. Hescheler, K. 98. Ueber Regenerationsvorgange bei Lumbriciden. II. Theil. Jena. Zeitschr., Bd. 31 (N. F., Bd. 24), Heft 3-4, pp. 521-604, Taf. 21-26. Holmgren, E. °99. Zur Kenntnis Spinalganglienzellen des Kaninchens und des Frosches. Anat. Anz., Bd. 16, pp. 161-171, 11 Textfig. Hunter, G. W. °98. Notes on the Finer Structure of the Nervous System of Cynthia partita (Verrill). Zool. Bull., Vol. 2, No. 3, pp. 99-115, 6 fig. in text. Julin, C. See BENEDEN, E. van, ET JULIN, C. Kolliker, A. von. °89. Das Aequivalent der Attraktionssphairen E. v. Beneden’s bei Siredon. Anat. Anz., Jahrg. 4, pp. 147-155, 3 Textfig. Kolster, R. :00. Ueber das Vorkommen von Centralkérpern in den Nervenzellen von Cottus scorpius. Anat. Anz., Bd. 17, pp. 172-173, 2 Textfig. Kostanecki, K., und Siedlecki, M. 97. Ueber das Verhaltniss der Centrosomen zum Protoplasma. Arch. f. mik. Anat., Bd. 48, pp. 181-278, Taf. 10-11. Lenhossék, M. von. °86. Untersuchungen iiber die Spinalganglien des Frosches. Arch. f. mik. Anat., Bd. 26, pp. 370-453, Taf. 15-16. Lenhossék, M. *95. Centrosom und Sphare in den Spinalganglienzellen des Frosches. Arch. fr mik. Anat., Bd. 46, pp. 345-369, Taf. 15-16. Lenhossék, M. von. *98. Untersuchungen iiber Spermatogenese. Arch. f. mik. Anat., Bd. 51, pp. 215-318, Taf. 12-14, 1 Textfig. 160 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Lenhossék, M. von. °98". Ueber Flimmerzellen. Verhandl. d. Anat. Gesell., Versamml. 12 (Anat. Anz., Bd. 14, Erganzungsheft), pp. 106-128, 3 Textfig. Lenhossék, M. von. 99. Das Microcentrum der glatten Muskelzellen. Anat. Anz., Bd. 16, pp. 334-342, 2 Textfig. Lewis, M. °96. Centrosome and Sphere in Certain of the Nerve Cells of an Invertebrate. Anat. Anz., Bd. 12, pp. 291-299, 11 fig. in text. Lewis, M. 98. Studies on the Central and Peripheral Nervous Systems of two Poly- chete Annelids. Proc. Amer. Acad. Arts and Sci., Vol. 33, No. 14, pp. 225-268, 8 pl. Leydig, F. 64. Vom Bau des thierischen Korpers. Bd. 1, Halfte, vi+ 278 pp. Tiibingen. Atlas. Mayer, S. b °72. Das sympathische Nervensystem. Stricker’s Handbuch der Lehre von den Geweben. Leipzig, Bd. 2, pp. 809-821, Textfig. 262-268. Also Eng. trans.: New York. 1872, pp. 767-776, fig. 285-291. McClure, C. F. W. 96. On the Presence of Centrosomes and Attraction Spheres in the Ganglion Cells of Helix Pomatia, with Remarks upon the Structure of the Cell Body. Princeton Coll. Bull., Vol. 8, No. 2, pp. 38-41. McClure, C. F. W. °97. The Finer Structure of the Nerve Cells of Invertebrates. Zool. Jahrb., Bd. 11, Anat. Abth., pp. 13-60, Pl. 2-3. Mead, A. D. 98. The Origin and Behavior of the Centrosome in the Annelid Egg. Jour. Morph., Vol. 14, No. 2, pp. 181-218, Pl. 16-19. Meves, F. 91. Ueber amitotische Kernteilung in den Spermatogonien des Salamanders und Verhalten der Attractionssphire bei derselben. Anat. Anz., Jahrg. 6, pp- 626-639, 11 Textfig. Meves, F. °95. Ueber cine Metamorphose der Attractionssphare in den Spermatogonien von Salamandra maculosa. Arch. f. mik. Anat., Bd. 44, pp. 119-184, Taf. 7-11. Meves, F. 795". Ueber die Zellen des Sesambeins in der Achillessehne des Frosches (Rana temporaria) und iiber ihre Centralkorper. Arch. f. mik. Anat., Bd. 45, pp. 133-144, Taf. 9. RAND: NERVOUS SYSTEM OF LUMBRICIDA. 161 Meves, F. °96. Ueber die Entwicklung der mannlichen Geschlechtszellen von Salaman- dra maculosa. Arch. f. mik. Anat., Bd. 48, pp. 1-83, Taf. 1-5. Meves, F. 98. Zellteilung. Anat. Hefte, Abth. 2, Ergebnisse d. Anat. u. Entwick., Bd. 8, pp. 430-542, 2 Textfig. Moore, J. E. S. 93. On the Relationships and Réle of the Archoplasm during Mitosis in the Larval Salamander. Quart. Jour. Mic. Sci., Vol. 34, pp. 181-197, Pl. 21. Moore, J. E. S. *95. On the Structural Changes in the Reproductive Cells during the Sper- matogenesis of Elasmobranchs. Quart. Jour. Mic. Sci., Vol. 28, pp. 275- 313, Pl. 13-16, 4 fig. in text. Niessing, G. *95. Zellenstudien. I. Arch. f. mik. Anat., Bd. 46, pp. 147-168, Taf. 5. Niessing, G. *99. Zellenstudien. II. Arch. f. mit. Anat., Bd. 55, pp. 63-110, Taf. 6. Pfitzner, W. °83. Beitrage zur Lehre vom Bau des Zellkerns und seinen Theilungserschein- ungen. Arch. f. mik. Anat., Bd. 22, pp. 616-688, Taf. 25. Pfitzner, W. 86. Zum morphologischen Bedeutung des Zellkerns. Morph. Jahbr., Bd. 11, pp- 54-77, Taf. 5. Rabl, C. °89. Ueber Zellteilung. Anat. Anz., Jahrg. 4, pp. 21-30, 2 Textfig. Rath, O. vom. *92. Zur Kenntniss der Spermatogenese von Gryllotalpa vulgaris Latr. Arch. f. mik. Anat., Bd. 40, pp. 102-132, Taf. 5. Rath, O. vom. "93. Beitrage zur Kenntniss der Spermatogenese von Salamandra maculosa. Il. Theil. Zeitschr. f. wiss. Zool., Bd. 57, pp. 141-185, Taf. 8-9. Rath, O. vom. °95. Zur Conservirungstechnik. Anat. Anz., Bd. 11, pp. 280-288. Rath, O. vom. *95*. Ueber den feineren Bau der Driisenzellen des Kopfes von Anilocra mediterranea Leach im Speciellen und die Amitosenfrage im Allgemeinen. Zeitschr. f. wiss. Zool., Bd. 60, pp. 1-89, Taf. 1-3. Rawitz, B. *95. Centrosoma und Attractionssphare in der ruhenden Zelle des Salaman- derhodens. Arch. f. mik. Anat., Bd. 44, pp. 555-579, Taf.°33. 162 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. Rawitz, B. °98. Untersuchungen iiber Zelltheilung. Arch. f. mik. Anat., Bd. 53. pp- 19-62, Taf. 2. Reinke, F. 94. Zellstudien. I. Theil. Arch. f. mik. Anat., Bd. 43, pp. 377-422, Taf. 22-24. Reinke, F. °95. Zellstudien. II. Theil. Arch. f. mik. Anat., Bd. 44, pp. 259-284, Taf. 19. Remak, R. 44. Neurologische Erlauterungen. Arch. f. Anat. Physiol. u. wiss. Med., 1844, pp. 463-472, Taf. 12. : Remak, R. 53. Ueber gangliédse Nervenfasern beim Menschen und bei den Wirbel- thieren. Monatsber. Akad. Wissensch., Berlin, 1853, pp. 293-298. Rohde, E. °98. Die Ganglienzelle. Zeitschr. f. wiss. Zool., Bd. 64, pp. 697-727, 5 Textfig. Schaffer, J. °96. Ueber einen neuen Befund von Centrosomen in Ganglien- und Knor- pelzellen. Sitzungsb. Akad. Wissensch., Wien, math.-naturw. Classe, Bd. 105, Abth. 3, pp. 21-28, 1 Taf. Schultze, H. °79. Die fibrillare Structur der Nervenelemente bei Wirbellosen. Arch. f. mik. Anat., Bd. 16, pp. 57-111, Taf. 5-6. Schultze, M. 71. Allgemeines iiber die Structurelemente des Nervensystems, Stricker’s Handbuch der Lehre von den Geweben. Leipzig. Bd. 1, pp. 108-136, Textfig. 17-30. Also Engl. trans.: New York. 1872, pp. 116-142, figs. 28-43. Schwalbe, G. °68. Ueber den Bau der Spinalganglien nebst Bemerkungen iiber die sym- pathischen Ganglienzellen. Arch. f. mik. Anat., Bd. 4, pp. 45-72, Taf. 4. Solger, B. °89. Zur Structur der Pigmentzelle. Zool. Anz., Jahrg. 12, pp. 671-673, 1 Textfig. Solger, B. 90. Nachtrag zu dem Artikel: ‘Zur Structur der Pigmentzelle.” Zool. Anz., Jabrg. 13, pp. 93-95. Solger, B. 91. Uecber Pigmenteinschliisse in der Attractionssphare ruhender Chro- matophoren. Anat. Anz., Jahrg. 6, pp. 282-284, 2 Textfig. | . ‘ { | ' i 24 RAND: NERVOUS SYSTEM OF LUMBRICIDA. 163 Spuler, A. °96. Beitrage zur Histologie und Histiogenese der Binde- und Stiitzsubstanz. Anat. Hefte, Abth. 1, Arbeiten, Bd. 7, pp. 115-158, Taf. 5-6. Stricht, O. van der. 92. Contribution a |’étude de la Sphére attractive. Arch. de Biol., Tome 12, pp. 741-763, Pl. 24. Van Beneden, E. See BENEDEN, E. VAN. Van der Stricht, O. See Srricut, O. VAN DER. Waldeyer, W. ’°88. Ueber Karyokinese und ihre Beziehungen zu den Befruchtungs- vorgangen. Arch. f. mik. Anat., Bd. 32, pp. 1-122, 14 Textfig. Walter, C. 763. Mikroskopische Studien tiber das Centralnervensystem wirbelloser Thiere. vilit 56 pp., 4 Taf. Bonn. Watasé, S. °93. Homology of the Centrosome. Jour. Morph., Vol. 8, No. 2, pp. 433- 443, 7 fig. in text. Zimmermann, K. W. 793. Studien iiber Pigmentzellen. I. Ueber die Anordnung des Archiplasmas in den Pigmentzellen der Knochenfische. Arch. f. mik. Anat., Bd. 41, pp. 367-389, Taf. 23-24. Zimmermann, K. W. °94. (Demonstration.) Verhandl. d. Anat. Gesell., Versamml. 8 (Anat. Anz., Bd. 9, Erganzungsheft), p. 245. Zimmermann, K. W. 98. Beitrage zur Kenntniss einiger Driisen und Epithelien. Arch. f. mik. Anat., Bd. 52, pp. 552-706, Taf. 27-29. VOL. XXXVII. — NO. € 5 164 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. EXPLANATION OF PLATES. All the figures were drawn with the aid of the camera lucida. Unless otherwise stated, all preparations were fixed in Flemming’s stronger chromic-osmic-acetic mixture. Owing to defective printing, centrosomes and chromatic granules sometimes appear as rings, instead of solid bodies. a. can.ali. chr. Cle. cl. mit. el. pr’f. c’so. cla. d. ec’drm. sé ec’drm’. gn. gn. nov. gn. Sw es. mb. mb’. mb. nl. Mu. Cre. ABBREVIATIONS. Anterior. mu. crc’. Alimentary canal. Chromatic mass (nucleus). mu. 1g. Cicatrix (cicatricular tissue). mu. lg’. Mitotic cell. Mass of proliferating cells. reg. pol. Centrosome. Cuticula. stmd. Dorsal. te. Old epidermis. trt. fbr. Regenerated epidermis. tu. Ganglion. tu’. Fundament of ganglia. a Fundament of brain. Mother-cell membrane. New cell membrane. Nuclear membrane. Old circular muscles. Regenerated circular mus- cles. Old longitudinal muscles. Regenerated _ longitudinal muscles. Polar region; cytoplasmic fundament. Stomodeum. Testis. Regenerated fibre tract. Sheath of nerve cord. Cut end of old sheath. In Figures 14-16, the * marks the point where the old epidermis ends and the regenerated epidermis begins. Rans. — Nerve-Cell Centrosome. PLATE 1. All the figures magnified 2000 diameters. Figs. 1-3. Cells from the subesophageal ganglion of a normal worm. Iron- hematoxylin. Fig. 4. Cell from the brain of the same worm. Figs. 5,6. Cells from a posterior ganglion of a normal worm. Acetic-sublimate, iron-hematoxylin. all = a! Aad Ranpb. — Nerve-Cell Centrosome. PLATE 2. All the figures magnified 2000 diameters. Figs. 7-9. Smaller cells from the posterior dorsal region of the brain of a normal worm. Iron-hematoxylin. Figs. 10-12. Cells from a regenerated brain after 34 days’ regeneration. These three cells lay side by side in the same section. Iron-hematoxylin. Fig. 13. Cell from a brain of 34 days’ regeneration. Iron-hematoxylin. ae ee a a ta, 2 gf a ae a — ; Panel 2 Ranp. — Nerve-Cell Centrosome. PLATE 3. Parasagittal section through the anterior end of a worm after 7 days’ re- generation. X25. Hermann’s plat.-acet.-osm., iron-hematoxylin. Parasagittal section through the anterior end of a worm after 16 days’ re- generation. X 25. Acetic-sublimate, iron-hematoxylin. Parasagittal section through the anterior end of a worm after 24 days’ regeneration. X 25. Acetic-sublimate, iron-hematoxylin. Spindle cells from the cicatrix of the worm from which Figure 14 is taken. (7 days’ regeneration.) X 2000. HOWE. del. 2p igeattgo game were ve al eg Oe: j i | ee Os ; i a* Ranp. — Nerve-Cell Centrosome. 5 ike} PLATE 4. Anterior end of the nerve cord of Figure 14, showing a mass of newly formed cells. (7 days’ regeneration.) X 160. Parasagittal section through the middle region of one lobe of the brain of anormal worm. X 160. Parasagittal section through the middle region of one lobe of a brain of 34 days’ regeneration. X 160. A dividing cell in the prophase and one in the metaphase (one pole cut away) from a brain of 34 days’ regeneration. X 2000. Iron-hema- toxylin. Metaphase from a brain of 34 days’ regeneration. 2000. Iron- hematoxylin. Metaphase from the anterior end of a cord after 34 days’ regeneration. x 2000. Gentian violet. Metaphase from a brain of 34 days’ regeneration. X 2000. Gentian violet. Telophase from a brain of 34 days’ regeneration. X 2000. Gentian violet. Telophase from the posterior dorsal region of a brain of 34 days’ re- generation. X 2000. Iron-hematoxylin. HW. de!. = Ranp. — Nerve-Cell Centrosome. PLATE 5. All the figures magnified 2000 diameters. Fig. 27. Telophase from a brain of 34 days’ regeneration. Gentian violet. Fig. 28. Telophase from a brain of 34 days’ regeneration. One chromatic mass cut away. Gentian violet. Fig. 29. Telophase from a brain of 34 days’ regeneration. Reconstructed from two sections. Above it, a young cell. Iron-hematoxylin. Fig. 80. Telophase from the anterior end of a cord after 34 days’ regeneration. Gentian violet. Fig. 31. Young cell (with several of the neighboring nuclei) from the posterior ventral region of a brain of 34 days’ regeneration. Jron-hematoxylin. Figs. 32a, 32). A pair of young sister cells from a brain of 34 days’ regeneration. The cells are in successive sections. Iron-hematoxylin. Fig. 33. Two pairs of young sister cells from a brain of 34 days’ regeneration. a and 8 are one pair. The two cells of the other pair lie one over the other. y is the upper of the two and the outline of the lower is indi- cated by 6. Gentian violet. Figs. 34, 35. Young cells from the anterior end of a cord of 34 days’ regeneration. Tron-hematoxylin. Fig. 36. A group of cells from the posterior ventral region of a brain of 34 days’ regeneration. At the centre of the group is a young cell. Iron- hematoxylin. Fig. 37. Young cell from the dorsal region of a brain of 34 days’ regeneration. Gentian violet. 28 i 30 ; ee Go, reg-pol, il mb ven, ‘| 33 cho. — q; : 31 (f> rn > ¥ 4 . ne ij ss t1,% T= ; & reg.pol. poetse” Von alia iam FS on Fe chAr ai? { “ee 6 ' 323 gS. 5 ete om ” ‘ 37 tote © oe 2 . +s gi 2 bd 4 . o*-% a se me : : & *< ea ine S, ° . + “* 4 t- ~ J . “ ” . . . é S. & ‘2 : . ¢ et a ae ee ~ d a ‘ Pa A ee f *e. e ary - 36 €". 7 & a No 2” = 9 “—~- ‘7 - fie» ‘ rie @ ig . = Ba. oi ~ > ‘ , te “~ ~Pe . ¥ r « - / i. “Pr ; ; e » 4 Stine’ 5 oe 4 b 3 ; “ , SASS * ¢ ce Pi - 7 e . e e., 4 ‘ —_— > ’ é . ; ' eo Ranpv. — Nerve-Cell Centrosome. Fig. 38. Fig. 39. Fig. 40. Fig. 41. Fig. 42. Fig. 43. PLATE 6. All the figures magnified 2000 diameters. Young cell (with several of the neighboring nuclei) from the anterior end of a cord of 34 days’ regeneration. Iron-hematoxylin. Small cells with resting nuclei from the anterior end of acord of 34 days’ regeneration. Each of the two cells has a centrosome. Iron-hema- toxylin. Cell with a greater mass of cytoplasm than in the cells of Figure 39; centrosome and radiations. From a brain of 34 days’ regeneration. Tron-hematoxylin. Cell, similar to that of Figure 40, from a brain of 37 days’ regeneration ; centrosome and radiations. JTron-hematoxylin. Three cells from the anterior end of a cord of 34 days’ regeneration. A centrosome with radiations may be seen near each of the two larger nuclei, on the side toward the greatest cytoplasmic mass. Iron-hema- toxylin. Two cells from the epidermis of anormal worm. The centrosome is at tlie deeper end of each nucleus. Iron-hematoxylin. Figs. 44, 45. Cells from regenerated epidermis, showing centrosomes and radia- tions. 16 days’ regeneration. Acetic-sublimate, Kernschwartz and safranin. en _— _ ~ — PLATE 6. NA tl. ser a Atay ene Ss } puer a “ o ve" s ry @ q if ' ™ Nei iA “ea Ff Ranv. — Nerve-Cell Centrosome. PLATE 7. All the figures magnified 2000 diameters. eee 46-52. Dividing or recently divided cells from the regenerated e a worm, after 11 days’ regeneration. Hermann’s plat.-acet.- hematoxylin. See pages 146-147 for a discussion of the "Ranp.— NERVE- CELL CENTROSONE. . § ire” fae a | vente a’G \ E x Aa Lo if q - = - ’ ‘ pe ( d : - - ’ va 1) #. . } r) 7 ‘ ial ai a ‘ f ai \ CO MMIMMMM DIS eA SAGAN io) a a a OY ey Hf, a Da by bb p 1 hy, iM 5 ma wet 7 i ; ah ; , ; Be i wiht , rT t hy ME BAY 4 et al ae | ' ly mu : wo i j i a ' , i | P) i f I | ¥ he : ) i if - are | } ’ Wi Lon J ¢ M ri } : 5 a] a ! Aw ‘ i! f i y : BHA ‘ } ’ 4 t J h : LI if by i} hi j j : ¥ ral eat ' i ¥ | 1 j : * i | i | +f ' ' 5 7 7 } i j : oF ! we } it i ‘ j uf \ + j ¢ 4 iT i " si in. i ‘ bey 44 i 4 7 i ; BINDING DOT. APRE Wal QL Harvard University. Museum a of Comparative Zoology H3 Bulletin v. 36-37 Biological & Medica! erials PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY