cilia on a slight elevation of the stomach wall ventral to the vesicular valve that closes the entrance of the oesophagus. The anus when opened extrudes a circle of cilia but these are perhaps not distinct from the ciliated lining of the intestine.

The process of digestion is somewhat as follows:

An Annelid larva was observed to pass between the longitudinal bands, which seemed to select the food, through the opened oesophagus and valve into the stomach where the paralyzed larva was soon resolved into refractive globules and setae. The patch of large cilia guarded the valve from contact with the coarser particles. The setae were collected and expelled in a bundle.

The transformation of the Tornaria into a Balanoglossus begins when the water-vessel measures about two-thirds of the distance from stomach to apex. The diverticula of the first pair of gill-slits then become noticeable by the active movement of their cilia. Migratory mesoblast cells become numerous. Vigorous contractions of the mesoblastic band accompany the enlargement of the water-vessel. The young Balanoglossus is less than one-half the size of the Tornaria and so opaque that sections are indispensable for following the internal changes. Hence it proved advisable to kill most of my specimens during or at short intervals after their transformation. Several which were kept for six weeks after changing to larval Balanoglossi appeared to have but two pairs of gill-slits. Though not old enough to identify, the larvae probably belong to B. Brooksii, a common species at Beaufort, whose casts are a characteristic feature of the sand flats at low tide.

Recent speculations regarding the phylogeny of the Enteropneusta have been modified by the prominence which is given to secondary larval characters. The basis for Metschnikoff's classification of the Euteropneusta with Echinodermata as "Ambulacralia" would be seriously impaired if it were proved that the points of similarity between Tornaria and Bipinnaria were the result of secondary adaptations to the same surroundings. But however this question may be ultimately decided, the homology of the water-vessel in the two larvae appears to be genuine and indicates relationship between the two groups.

Huxley includes Balanoglossus with the Tunicates in his division of "Pharyngopneusta," and draws attention to "the extraordinary similarity in the structure of the perforated pharyngeal sac in the larvae of Tunicates and of Balanoglossus."

The larva of B. Kowalerskii as elucidated by Bateson shows additional points of resemblance to the Chordata, especially to Amphioxus in the notochord, atria, etc.

A closer study of the transformed Tornaria in regard to its internal changes and later development would show how far it confirms these affinities of B. Kowalerskii. But whether Tornaria or Bateson's larva is to be considered as more nearly representing the ancestral type, there is still the possibility that each may possess certain phylogenetic characters that have become obsolete in the other.

The Farasitic Cuninas of Beaufort. By H. V. WILSON. There have been described three classes of parasitic Cuninas or (Haeckel) Narco-medusae. The first group includes the Cuninas found in the subumbrellar cavity of Turritopsis (Cunoctantha octonaria H.). To the second belongs the eight tentacled Cunina that develops in the stomach of certain Geryonidae, and in the third group come those Cuninas found in the stomachs of other Cuninas. In the last division the host and parasite agree in some cases and differ in others, in the number of tentacles. The parasitism of the contained medusae is beyond question in the first two groups, and though not definitely settled is very probable in the third group. The Turritopsis Cunina was discovered long ago in the Charleston harbor by McCrady and was re-discovered by Dr. Brooks in 1882 at Beaufort. I again observed it this summer. The Cunina found in the Geryonidae (Carmarina, Geryonia) has been observed on our coast in but a single instance, and then Fritz Müller thought the little Cunina had been swallowed by the Liriope. During July and August I found a number of Liriope (the common Geryonid at Beaufort) with Cuninas in the gastric cavity. My youngest stage was a thin two-layered sac with a single mouth opening. The next was a similar sac with two tentacular protuberances by which it was attached, and three mouths. In the next stage at the point of each mouth, a medusa bud was developed. These stages differ slightly

from Metschnikoff's description, in that there is no sign of the central pseudopodial cell, though in Metschnikoff's Cuninas it persisted up to a stage considerably later than my third. Another point of difference is, in the American Cunina the medusa buds are sprouted out over the general surface of the sac, there being no part of the sac specialized as a stolon from which a close agglomeration of buds could be produced, as Metschnikoff figures. The Cuninas parasitic in Cunina have hitherto been observed only in the Mediterranean. Through August, I got in the tow-net a number of eight tentacled Cuninas differing from Cunoctantha octonaria only in being perfectly transparent, and in having a much smaller mouth with no pigment round it. No parasites were observed in this medusa until the last of August, when I obtained a single specimen with half a dozen little Cuninas in its stomach. The parasitic Cuninas were lying free in the stomach, were perfectly transparent, and had twelve to fifteen tentacles.

Segmentation of the Egg and Formation of the Germ Layers of the Squid (Loligo Pealii). By A. T. BRUCE. The segmentation of the egg and formation of the germ layers of the Cephalopoda have been observed by several well-known investigators. Some points, however, relating to the origin of layers and, particularly, of the inner layer or endoderm are still doubtful.

Last summer while enjoying the facilities for study offered by the table of the Johns Hopkins University at the U. S. Fish Commission Laboratory at Wood's Holl, Mass., a good opportunity was offered for the study of the development of the common Squid (Loligo Pealii).

Two Squids kept in an aquarium laid eggs on the night of July 3d. My attention was called to this by Prof. John A. Ryder.

The earliest stage was studied by means of sections of eggs hardened the next day. Subsequent stages up to the time of hatching were also preserved and studied. For the study of segmentation and formation of germ layers the sectional method on many accounts is preferable to superficial observation. This method demonstrated in a very satisfactory manner the origin of the germ layers.

The youngest eggs studied had the protoplasm segregated at one pole forming a germinal disk. This disk in eggs which probably had not been laid twelve hours before hardening was already segmented.

The

The segments divided the disk into a number of rectangular areas. cleavage planes between the segments extended through the entire thickness of the germinal disk.

In older eggs which were probably hardened about twenty-four hours after deposition the germinal disk by a process of delamination had become divided into two layers. This delamination does not affect the entire surface of the germinal disk.

One axis of the disk probably corresponded to the long axis of the embryo was marked by an area but one cell thick. This area or axis does not however extend the entire length of the germinal disk. Transverse sections beyond the poles of the axis are two cells thick throughout. Sections through the axis are two cells thick at the lateral margins between which the area of the axis is a single cell thick. The mesoderm or deeper layer of the disk is consequently in the region of the axis separated into two masses or bands.

In the region of the axis, and probably anterior and posterior to it, also certain cells were observed partially separated from the mesoderm cells. These cells from their spindle shape and oval nuclei are readily recognized as the endoderm cells which form the inner coating of the yolk sack and digestive tract in more advanced embryos.

No nuclei were observed in the yolk at this or subsequent stages. It seems probable, then, that no endoderm cells originate spontaneously in the yolk, but that all are derived by growth and division from the earliest endoderm arising as described from beneath the mesoderm. . . . .

In eggs about thirty-six hours old the mesoderm bands had become further divided and were two cells thick. The endoderm at this stage by the growth and division of its cells formed a thin stratum over the inner concave surface of the germinal disk resting upon the yolk.

In eggs about sixty hours old the ectoderm and endoderm had extended so much beyond the area comprising the original germinal disk as to enclose nearly all the yolk. The mesoderm at this stage did not cover much more

than half the egg, its terminal margin forming a line a little below the equator of the egg. Consequently the egg at this stage may be divided into two hemispheres: an upper hemisphere covered by ectoderm, mesoderm, and endoderm, and a lower hemisphere covered by ectoderm and endoderm. The upper hemisphere is the embryonic area, and the lower hemisphere is the area of the yolk sack. The line separating the hemispheres and marking the terminal edge of the mesoderm was marked externally by a slight depression.

At this stage a slight eminence in the middle of the embryonic area marked the first trace of the mantle. The separation of the mesoderm into two bands was apparently not retained at this stage. The ectoderm was not sharply marked off from the mesoderm. The development of organs belongs to later stages. The embryos hatched on the 18th of July.

Notes on the Flora of Abaco and adjoining islands. By F. H. HERRICK.

The flora of Abaco, so far as we observed it, differs widely from that of the smaller islands or keys which form its inner reef. The whole island is covered from end to end by a rank forest growth, composed of a very great variety of trees and shrubs.

The Pine, probably Pinus Cubensis, is the principal forest tree, and strikingly resembles the Pinus taeda or Loblolly pine of the Southern States. The former and the Red Cedar, Juniperus Virginiana, were the only conifers which we saw. Cedar prevails in the vicinity of certain swamps towards the north-eastern extremity of the island. Above this point the pine reaches its greatest size, attaining a height of 80 feet. Other common and important trees are the Mastic, "Dogwood," White Torch, Bullet, and "Poison" woods. The last is probably a species of Rhus. It has a scaly light brown bole, greenish flowers and pinnate leaves, the juice of which stains the skin black, and produces an inflammation, similar to that caused by our common Poison Ivy, Rhus Toxicodendron. The low, swampy parts of all the islands are almost exclusively occupied by the Mangrove, Rhizophora Mangle, which forms dense, impassible thickets. This shrub had passed its flowering by the first of June, and the peculiar, elongated seeds were in all stages of development.

The most interesting feature of Abaco, however, is the great variety and beauty of its epiphytic plants,-Orchids, Ferns, and Bromeliads. In some places every tree and shrub, however insignificant in size, is ornamented by one, often by several species of these attractive plants.

The Keys, about a dozen of which we visited, were characterized by the absence of the conspicuous forest trees seen on Abaco. Many of the smaller islands, which possess a very scanty soil, are carpeted with a heavy matting of vines, twining plants, and a few grasses and small shrubs. Among these we find "Rock Samphire," or "Turtle grass" (Gomphrena), running prostrate plants with turgid stems and leaves: Convolvulii or Morning Glories of several species, one of which has very large reddish flowers: Carnavalia, a pea vine having long rope-like stems and small pink blossoms. Echites suberecta, which belongs to a genus of tropical plants, is a very common and showy climber, often completely covering the lower shrubbery. It is distinguished by its large lemon-colored flowers, shining leaves, and milky juice. A beautiful maroon-colored passion flower, Passiflora cupraea, was found growing on Pawpaw Key.

Joe's Key, off Little Abaco, differed somewhat from any of the islands which we visited. It contained a small grassy meadow drained by a mangrove swamp on either side, a grove of silver-topped palms; a sandy beach on the leeward side, and high cliffs facing the sea. Uniola paniculata, SpikeGrass (common along our Southern sea-coast), with several tall sedges was growing along the sand beach, and also a delicate plant, Sabbatia gracilis, which has large rose-colored flowers.

The Pancratium, or "West Indian Lily," a member of the Amarillis family, a plant the bulbs of which are regularly imported by florists, grows commonly along the sandy margins of all these islands. Its flowers are large, pure white, and remarkably fragrant.

The larger keys are generally covered at all available points by an almost impenetrable growth of shrubs and small trees, including several species of palm, and a large bushy cactus (Opuntia). Perhaps the most characteristic shrub is Rhacicallis rupestris, called "Seaweed" by the natives. It is pros

trate or partially erect, and has dark green sprayey foliage, and minute saffron flowers. It is usually confined to the rocks along shore. The Conocarpus or "Button-wood," often the only shrub of considerable size on the smaller islands, everywhere abounds, and is conspicuous by its silvery foliage. Besides these we constantly meet with the Wild Sapodilla, Sapota Achras, the fruit of which resembles a small rusty apple; Genipa clusifolia or "Sevenyear Apple," which has fragrant cream-colored flowers and a hard green fruit about the size of a walnut; Borrichia arborescens, "Lavender,” a yellow composite with tall brittle stems and hoary leaves, which are used for medicinal purposes.

Green Turtle Key probably affords a more diversified flora than any of the islands of equal area. Its habitation by man has led to the introduction of various foreign fruits and a host of common weeds. Some fine specimens of the Cocoanut Palm may be seen here. This palm also flourishes along the shores of Abaco and on some of the less exposed islands. A small grove of these trees on Allons Key was completely prostrated by the great hurricane of September 7, 1884. Such fruits as the Banana, Orange, Lemon, Shaddock, Sapodilla, Mango, Mammee, Custard Apple or Papaw, Hog Plum, Tamarind, Date Palm, Fig (the two last represented by a few specimens only), have been introduced and are now growing spontaneously over the island.

The Coccoloba or "Sea Grape" forms a low contorted shrub, or as frequently attains the height of a small tree. It is remarkable for its thick, nearly circular leaves, and long pendulous racemes of green flowers or fruit. By far the most striking herb on the Key just mentioned is the Manilla plant, a species of Agave, whose solitary flower stalk attains a height of upwards of 20 feet, and forms a prominent feature in the landscape.

The Pine-apple, Ananassa sativa, is grown extensively on Abaco in clearings along the shores, and is the only fruit exported from this portion of the Bahamas.

List of Plants from Abaco Island, Bahama. By D. C. EATON and W. A. SETCHELL, from collections made by F. H. HERRICK.

Note.-A small collection of plants from portions of Abaco and several of its adjoining islands was made last June, when the Marine Laboratory was stationed at Green Turtle Key.

Prof. D. C. Eaton of Yale College, New Haven, Ct., very kindly consented to examine this collection, and to him and to Mr. W. A. Setchell of the same institution, we are indebted for the determination of nearly all the plants included in the list below.

We were unfortunately unable to preserve suitable specimens of a great many plants, and such as we obtained were often too fragmentary to admit of certain identification.

About forty natural orders or families are represented by the sixty odd species given in the list. Those plants marked by a †, about a third of the whole number, are not known to occur in the Southern States or on the Florida Keys, and are probably peculiar to the West Indies. Plants found only on Abaco Island are thus indicated, and those observed only on Green Turtle Key are marked by the letters G. T. Local popular names are placed in quotation marks.

FLOWERING PLANTS.

1. Agave sps (?). "Manilla Plant."

2. Alternanthera flavescens, Moquin.

3. Argemone Mexicana, L. Prickly Poppy. (G. T.)

4. Artemisia vulgaris, L. Common Mugwort. (G. T.)

5. A. hispida, Pursh.

nium."

6. Asclepias paupercula, Michx. Milkweed.

F. H. H.

"Bastard Gera

Observations on the Nervous System of Insects and Spiders and some Preliminary Observations on Phrynus. By A. T. BRUCE.

From a careful study of the nervous system of some insects (Acridium and Thyridopteryx) and of spiders it appears that the supraœsophageal ganglion of insects and Arachnids is distinctly double. Each ganglion of the ganglionated chain of insects and spiders is closely invested by a sheath of flattened mesoblast cells, which may be described as the perineurium. The perineurium in insects also invests the transverse and long commissures between ganglia and the nerve trunks.

The perineurium extends between adjacent ganglia forming a close sheath for each ganglion. Consequently adjacent ganglia are separated by ingrowths of perineurium. The perineurium originates in part as a median ingrowth along the median ventral line. The mode of origin of this ingrowth was described in Circular No. 49, p. 85. All the median ingrowth becomes perineurium except portions of it between adjacent pairs of nerve ganglia where it persists as columns of cells extending to the dorsal limit of the nervous system.

It is apparently by the extension and union of columns such as these that the endosternite of Limulus and Arachnids is formed. The nerve ganglia are first solid masses of cells. By the breaking down of cells in the centre of each ganglion the commissural or punctsubstance is formed.

By the extension of the punctsubstance laterally and internally the cross or transverse commissures are formed, and by a similar extension of the punctsubstance anteriorly and posteriorly the long commissures are formed. The central portion of each ganglion consisting as described of punctsubstance is continuous externally and laterally with the nerve fibres. It is by the breaking down of the cells of the originally solid nerve cord that the nerve fibre is formed. All stages in the formation of commissures and fibres were not observed.

In comparing the supracesophageal ganglion of insects and spiders with other ganglia of the chain it appears that the supraœsophageal ganglion of insects and spiders is distinctly double. Each division is serially homologous with a ganglion pair for succeeding body somites. In both insects and spiders the anterior division of the supraœsophageal ganglion is distinctly separated from the posterior division by the ingrowth of the investing peri

neurium. In insects (Thyridopteryx) the columns of cells described as existing between other pairs of ganglia exists between the two divisions of the supracesophageal ganglion.

In Thyridopteryx the anterior division of the supracesophageal ganglion is marked externally by an incision of ectoderm as are the succeeding body segments. There can be no doubt then but that the brain of insects and spiders consists of two pairs of ganglia serially homologous with other pairs of ganglia.

The anterior division of the supracesophageal ganglion of insects is the pair of ganglia belonging to the antennal somite. The posterior division of the supracesophageal ganglion of insects is the ganglion of the segment of the upper lip. The upper lip is a paired structure in Thyridopteryx and Acridium. It receives nerves from the ganglia of its somite. The posterior division of the supracesophageal ganglion of insects by its backward extension forms most of the circumoesophageal commissure, consequently the labrum is innervated from the commissure as is the second antenna of crustacea. The anterior division of the spider's brain probably innervates the rostrum. The lateral parts of the rostrum form downward projections on each side and correspond in position to the antennae of insects. These antennal (?) projections are well shown in transverse sections. The posterior division of the spider's brain innervates the labrum, which is distinctly grooved in the median line and has the appearance of a paired structure. The mandibular ganglion of spiders lies below the posterior division of the supracesophageal ganglion, and is partly overlapped by it laterally.

There appears, then, to be a special homology existing between the labrum in spiders and insects and the recurrent antenna of crustacea. There is also a correspondence between the antenna of insects, the first antenna of crustacea, and the lateral parts of the rostrum in spiders.

The eyes of insects (Acridium and Thyridopteryx) appear first as thickenings of ectoderm on the antero-lateral parts of the head. In Thyridopteryx this area becomes further specialized to form distinct optic elements or ocelli, formed, primarily as solid areas or cones. Subsequently, in all probability, by the breaking down of the central cells of the cone, the punctsubstance or granular substance of the ocellus is formed, surrounded by a layer of cells which, at the periphery, is double. These elements of the eye, or ocelli, in Thyridopteryx are, in the advanced embryos, separate from each other and not connected with the central nervous system. The development of the eye in the grasshopper (Acridium) was not followed beyond the very earliest stages.

Some observations were made in a pedipalp (Phrynus) with embryos brought from the Bahama Islands by Dr. Orr. The ventral surface of the thorax of the adult pedipalp closely resembles the ventral surface of Limulus. A striking point of likeness is the existence of a curved process on the inner side of the coeal joint of the last thoracic appendage, corresponding to a similar process on the coeal joint of the last appendage of Limulus. Traces of similar processes also appear in the fourth and fifth thoracic appendages.

Another interesting likeness to Limulus is the existence of a paired structure posterior to the last pair of thoracic appendages, corresponding in position to the chiloria or metastoma of Limulus. Each of the three last pairs of thoracic limbs possess episterna at their bases, between which is a median dark-colored sternum, that for the last pair of thoracic limbs being double. The metastoma of the Pedipalp has anterior to it a paired structure, which is, perhaps, a paired sternite, and at its base a paired structure which is an episternite.

These points of likeness together with the many familiar ones, and the involution of the abdominal feet of the spider to form the lung book described by me in a recent number of the American Naturalist, make the relationship of Limulus and the Arachnids very close.

The embryo pedipalp has an amnion like that described by me for spiders. On the coeal joint of the 4th appendage a sense organ was very conspicuous in the young pedipalp. The hypodermic cells of that joint become columnar, unlike other hypodermic cells, which are of irregular contour. These elongated cells are continued externally to form filaments, several of which enter a single pair, which is the external part of the sense organ.

48

[No. 54.

I. American Journal of Mathematics.

The publication of this journal commenced in 1878, under the editorial direction of Professor Sylvester. It is now conducted by Professor Simon Newcomb, as Editor, and Dr. T. Craig, as Associate Editor. Eight volumes of about 400 pages each have been issued, and the ninth is in progress. It appears quarterly, in the quarto form. Subscription $5 per year. Single numbers $1.50.

II. American Chemical Journal.

This journal was commenced in 1879, with Professor Remsen as Editor. Seven volumes of about 450 pages each have been issued, and the eighth is in progress. It appears bi-monthly. Subscription $3 per year. Single numbers 50 cents.

III. American Journal of Philology.

The publication of this journal commenced in 1880, under the editorial direction of Professor Gildersleeve. Six volumes of about 570 pages each have been issued, and the seventh is in progress. It appears four times yearly. Subscription $3 per volume. Single numbers $1.00.

IV. Studies from the Biological Laboratory.
(Including the Chesapeake Zoological Laboratory.)

The publication of these papers commenced in 1879, under the direction of Professor Martin, with the assistance of Dr. W. K. Brooks. Two volumes of about 500 pages octavo, and 40 plates each, have been issued, and the third is in progress. Subscription $5.00 per volume. (See page 41 of this Circular.)

V. Studies in Historical and Political Science. The publication of these papers was begun in 1882, under the editorial direction of Dr. H. B. Adams. Four series of twelve monthly parts each have been issued, and the fifth is in progress. Subscription $3.00 per volume. Extra volumes, bound in cloth, are also issued from time to time.

VI. Johns Hopkins University Circulars. The publication of the Circulars began in December, 1879, and fifty-four numbers have since been issued.

The Circulars are published at convenient intervals during the academic year for the purpose of communicating intelligence to

the various members of the university in respect to work which is here in progress, as well as for the purpose of promulgating official announcements from the governing and teaching bodies. During the current academic year, circulars will be issued in successive months, to be followed at the close of the year by an index.

Although these circulars are designed for the members of the university, they have frequently been called for by institutions and libraries at a distance, and also by individuals who are interested in the literary and scientific activity of this university. Subscriptions and exchanges are therefore received. Subscription $1.00 per year. Subscribers to the Circulars will also receive the Annual Register and the Annual Report of the University.

VII. Annual Report.

Presented by the President to the Board of Trustees, reviewing the operations of the University during the past academic year.

VIII. Annual Register.

Giving the list of the officers and students, and containing detailed statements as to the regulations and work of the University.

Announcements of proposed lectures, courses of instruction, etc., appear in the University Circulars, or are separately issued as Programmes from time to time.

The following separate publications have also been recently issued by the Publication Agency of the University:

Photograph of the Normal Solar Spectrum. Made by Professor H. A. Rowland. Set of seven plates unmounted $10, mounted $12; single plates unmounted $2, mounted $2.25.

Reproduction in Phototype of a Syriac MS. with the Antilegomena Epistles. (Including seventeen phototype pages, each on a separate leaf, besides additional pages containing descriptive and explanatory matter.) Edited by Dr. I. H. Hall. In paper $3; in cloth $4. Selected Morphological Monographs. Edited by Dr. W. K. Brooks. In cloth $8. (See announcement on page 41 of this Circular.)

Communications in respect to exchanges and remittances may be sent to the Johns Hopkins University (Publication Agency,) Baltimore, Maryland.

DERSLEEVE.

[Abstract of a paper read before the University Philological Association, December 3, 1886].

The specimen presented in this paper consisted of a new translation of the speech of Aristophanes with some remarks on the succession and connexion of the various discourses on Eros. The following extracts will show the drift; the detailed work is reserved for publication elsewhere.

The first speaker is the sucking sophist Phaidros-who shamelessly suggests the theme on the ground of its novelty, in reality, as we may gather from the Platonic dialogue which bears his name, because he has carefully prepared a little discourse on the subject, of which he is only too glad to deliver himself before so choice a company. A facile friend lends himself to the scheme, the conversation is brought round to the desired point and Phaidros takes the lead. The sucking sophist is succeeded by the riper sophist Pausanias and Pausanias is followed by the physician Eryximachos, who like most physicians has a pet philosophy of his own and tries to better the theories of others by his vast experience of the healing art. These three speeches stand nearly related to each other. Each makes an advance on its predecessor. The analysis which is crude and implicit in Phaidros becomes clear and explicit in Pausanias Phaidros has not yet passed from the class of rhetoric into the class of logic. He has no real power of analysis. The dichotomy which announces itself at once in the speech of Pausanias gets itself done very awkwardly in the speech of Phaidros. Phaidros gives us a mythological chaos from which Earth and Eros emerge, but there is only one Eros. There is no dichotomy except so far as the myth shows a dichotomy, the dichotomy of sex, the dichotomy of age-Alkestis and Admetos, Orpheus and Eurydike, Achilles and Patroklos. The lover is exalted above the beloved because of self-sacrifice, the beloved is exalted above the lover because of conscious agency, which is a nobler thing than irresistible impulse. Inconsistency does not trouble the rhetorical soul of Phaidros. There is a sophistic toying with mythology, a sophistic criticism of the poets, a sophistic emphasis as to the conclusions reached, a sophistic forging of the answer.

Pausanias follows Phaidros closely, and the absence of an interlude indicates the immediate connexion. Acknowledged and enthusiastic lover of Agathon, Pausanias goes deeper than Phaidros, even to the dividing asunder of Eros and Eros. The dichotomy is announced at once and we hear throughout the sophistic distinguo. Love is not one, and we must first establish

[PRICE, 10 CENTS.

what manner of Love is to be praised and then praise that Love in a manner worthy of the god. The opening seems to be as fair as Sokrates himself could have made it, but as we go on we find that Pausanias does not reach his dichotomy by dialectic; he simply posits what he wants. The speech is spun out to almost intolerable length so that we welcome the poor pun Пavoavíov dè πavoaμévov as a relief from the distinguo's here and the distinguo's there. And after all the conclusion is rough and forced, for he stops when he thinks he has made his point, which he never loses sight of through all his analysis. All that Pausanias wishes to do is to recommend his relation to Agathon as something noble and elevating. That is the real meaning of all his distinguo's and there is no honesty in him. Pausanias has fooled Schmelzer, the latest interpreter of the Symposium, Aristophanes he could not fool.

Eryximachos, the physician, accepts the dichotomy, but in accepting it, like all pedantic natures, he tries to better the instructions of his predecessor. He felt a certain incompleteness in the speech of Pausanias, but whereas we should say that Pausanias has not fairly carried out his rhetorical scheme, he says that Pausanias has not pushed his dichotomy far enough. To put what Eryximachos says into modern terms, it would be somewhat like this: "Love is not a mere craving for sympathy, pure or impure, heavenly or pandemic, corporeal or spiritual. Love is nature's abhorrence of a vacuum." And this is universally true. Love manifests itself in everything in nature. It is not merely the passion of the soul towards beauty, not merely the appetite of the human body. It is the appetence of all animals, all plants, all beings. The god is great and marvellous and universal, because the relation of plenum and vacuum is great and marvellous and universal. This Eryximachos sets forth with great selfconfidence-in his own peculiar province of medicine. When he comes to music, the application of his doctrine is much more constrained and hobbling, and he is evidently glad to get back to more familiar ground, where his mechanical theory has somewhat easier play.

Eryximachos shows at once the dishonesty of the sophist and the dogmatism of his profession in trying to make good his pedantic correction of Pausanias, and by the use of fine phrases works himself up to the belief in his own triumphant cleverness. With a self-satisfied smirk he challenges Aristophanes to fill any void he may have left, and at the same time calls attention to the fact that he has overcome Aristophanes' hiccough.

Thus far the analysts. Aristophanes redintegrates. He does not accept the challenge of Eryximachos to better his discourse by correcting and complementing it, as Eryximachos had done for the speech of Pausanias, as