Faivre very much bettered, demonstrate that a hexapod may live for months without a brain, if the subæsophageal ganglion, or better, ventro-cerebron, is left intact, just as a vertebrate may live without its cerebrum. Faivre long ago showed that this ventro-cerebron is the seat of the power of co-ordination of the muscular movements of the body. Binet has shown that the brain is the seat of the power directing these movements. A debrainedhexapod will eat when food is placed beneath its palpi, but it cannot go to its food even though the latter be but a very small space removed from its course or position. Whether the insect would be able to do so if the mushroom bodies only were destroyed, and the antennal lobes, optic lobes, and the rest of the brain were left intact, is a question that yet remains to be answered. In Binet's experiments neither olfactory nor visual stimuli can be transformed into motor impulses. Were it possible for them to be so transformed, my studies to be noted in a moment cause me to think that Binet's results would be very materially altered.

Now, as to my studies. During the winter just past with no little patience I endeavored to apply the bichromate of silver method to a study of the brain and general nervous system of the common honey bee, the more detailed result of a portion of which will be published a little later. The endeavor was rewarded by a considerable degree of success, the main facts being determined, though there are many details left for future studies. Others have tried to employ the same general method, but owing to a lack of proper store of patience or to their setting about the task wrongly have failed. Among them must be counted Binet ('94), with whom, however, there seems to be a defect in the conception of both the Golgi and the Erlich methods. For he sets the former aside as inconstant, uses the latter, without, however, apparently obtaining any very good results. He complains that preparations by the Erlich method (and the Golgi method might be included) leave out many details, and never seems to think that a sufficient number of preparations will supply those details and thus allow the whole to be determined. This is the more unfortunate, since his dependence upon the old methods has led him to give detailed

importance to phenomena that are relatively unimportant, and has resulted in a somewhat misty conception of the structure of the hexapod ventral nervous system.

One of the very first things that an impregnation of bee brains with bichromate of silver enabled me to make out was the structure of the mushroom bodies with their cells. These cells stand out in sharp contrast to all other nerve cells known, though they recall to some extent the cells of Purkinje in the higher mammals. Each of the cells contained within the fibrillar cup seeds a nerve process into the later, where it breaks up into a profusely arborescent system of brahchlets, which often appear with fine, short, lateral processes, such as are characteristic of the dendrites of some mammalian nerve cells. Just before entering the fibrillar substance a fine branch is given off that travels along the inner surface of the cup along with others of the same nature, forming a small bundle to the stalk of the mushroom body, down which it continues until it reaches the origin of the anterior and the inner roots mentioned at the beginning of the paper. Here it branches, one branch continuing straight on to the end of the anterior root, while the other passes to the end of the inner root. Throughout its whole course the fiber and its two branches are very fine. Nearly the whole stalk and nearly the whole of each root is made up of these straight parallel fibers coming from the cells within the cup of the mushroom bodies. What other fibers there are enter these bodies from the side, and branch between the straight fibers very much as the dendrites of the cells of Purkinje branch among the parallel fine fibers from the cells of the granular layer in the mammalian cerebellum. These fibers are of the nature of association fibers.

From the olfactory or antennal lobe, from the optic ganglia there are tracts of fibers that finally enter the cups of the mushroom bodies as shown by Viallanes and by my studies with the Golgi method and also with a Formol-copper-hæmatoxylin method of staining. Besides these tracts the Golgi method has enabled me to make out another tract, unknown before, passing down the hinder side of the brain from the cups to the region above the esophagus, where it bends forwards and comes in

contact with fibers from the ventral cord, which exists, although Binet was unable to discover any "growth of fibers connecting the cord with the brain."

The fibers entering the cups from the antennal lobe, the optic ganglia, and the ventral region, spread out and branch among the arborescent endings of the mushroom body cells.

Fig. .--A. An "intellective" cell from the mushroom body. n, neurite; d. dendrite; a.r., anterior branch of the neurite; i.r., inner branch of the neurite. B. Mushroom body of right side from above. The outer one, m.b, viewed in section; the inner one is cut off, leaving the stump of the stalk st. a.r., anterior root; i.r., inner root; m.b., cup.

The fibers branching among the parallel fibers of the roots and the stalk lead off to lower parts of the brain, connecting with efferent or motor fibers, or with secondary association fibers, that in their turn make such connections. This portion of the circuit has not been perfectly made out, though there seems to be sufficient data to warrant the assumption just made.

Such fibers existing as described there is then a complete circuit for sensory stimuli from the various parts of the body to the cells of the mushroom bodies. The dendritic or arborescent branches of these cells take them up and pass them on out along the parallel fibers or neurites in the roots of the mushroom bodies as motor or other efferent impulses.

This, however, is not all. For there are numerous fibers evident in my preparations, the full courses of which I have not been thus far able to determine, but which are so

situated as to warrant the inference that they may act as association fibers between the afferent fibers from the antennæ, optic ganglia, and ventral system and the afferent fibers. There is then a possibility of a stimulus entering the brain and passing out as a motor impulse without going into the circuit of the fibers of the mushroom bodies, or, in other words, a possibility of what may be compared to reflex action in higher animals.

It appears then that the supposition of Dujardin is well supported by the finer structure of the hexapod brain. For it is evident from the details known since the publication of Flögel's paper, that the cells composing the mushroom bodies have been very highly differentiated in some of the hexapods, and this in just those forms living the most complex lives. No such bodies are to be found in the lower forms. I have never seen them, nor any indication of them, in the Thysanura, Chilopoda, Scolopendrella, the Pauropoda and other Myriapoda, nor in any of the Crustacea that I have thus far examined. Without doubt an application of the Golgi or methylen blue methods would reveal elements in some these forms that might be compared with the cells of the mushroom bodies; but they would probably be found not so completely differentiated from other fibers as they are in the honey bee and other Hymenoptera. It may be mentioned that one does not recognize such cells in the cray-fish and the crab as figured by Retzius and Bethe. And it scarcely need be said that no such elements are shown in Retzius' figure of the brain of Nereis.

BIBLIOGRAPHY.

Bellonci, '82. Intorno alla struttura e alle connessioni dei lobi olfattori negli artropodi superiori e nei vertebrati. Reale Accad. d. Lincei. (From Cuccati.)

Berger, '78. Untersuchungen über den Bau des Gehirns und der Retina der Arthropoden. Arb. d. Zool. Inst. Wien u. Triest., I, 173-220.

2 St.-Remy ('90) describes mushroom bodies as occuring in Scutigera, which if homologous with the mushroom bodies of Hexapoda, is in accordance with Dujardin view.

Bethe, '95. Studien über das central nerven system von Carcinus mænus nebst ein neues Verfahren der Methylenblaufixation. Arch. f. Mikr. Anat., XLIV, 579–622.

Binet, '94. Contribution à l'étude du system nerveux sousintestinal des insectes. Journ. l'Anat. et Physiol., XXX, 449– 580.

Brandt, '76. Anatomical and Morphological Researches on the Nervous System of Hymenopterous Insects. Ann. Mag. Nat. Hist., (4) XVIII, 504–6.

Cuccati, '88. Uber die Organization de Gehirns der Somomya crythrocephala. Zeit. f. wiss. Zool., XLVI, 240–69. Diehl, '76. Die Organization des Arthropoden Gehirus. Zeit. f. wiss. Zool., XXVII, 488-517.

Leydig, '64. Vom Bau des tierischen Körpers. (From Viallanes.)

Flogel, '78. Ueber den einheitlichen Bau des Gehirns in den Verschiedenen Insekten Ordunung. Zeit. f. wiss. Zool., XXX, Supplement, 556-92.

Forel, '74. Les Fourmics de la Suisse.

Rabl-Ruckhard, '75. Studien über Insekten gehirne. Reichert und Du Bois Raymond's Arch. f. Anat., 488-99.

Packard, '80. The Brain of the Locust. Second Rept. U. S. Ent. Com., pp. 223–242.

Retzins, '90. Zur Kenntnis des Nervensystems der Crustaceen. Biol. Untersuch., N. F., I, No. 1.

Retzius, '95. Zur Kenntnis des Gehirnganglion und des sensiblen Nervensystems der Polychäten. Biol. Untersuch., N. F., VII, No. 2.

Saint-Remy, '90. Contribution á l'étude du cerveau chez les Arthropodes trachéátes. Lacaze Duthiers' Arch. d. Zool. Exper. et gén. (2) V sup. 4th mém.

Dujardin, '50. Mémorie sur le système nerveux des insectes. Ann. Sci. Nat., (3) XIV, 195–206.

Viallanes, '87. Le cerveau de la Guêpe. Ann. Soc. Nat., (7) II, 5-100.

IV.

Viallanes, '88. Le cerveau du criquet. Ann. Sci. Nat., (7)