connective filaments the middle one alone remains visible and receives a certain number of fibres from the ascending fasciculus. To complete our description let us glance at the series of transverse sections. In Figure 19 the two crural lobules have not yet united and are not yet merged into the dorsal-ventral lobe. This junction does not take place until we come to Figure 20. Here, at this level, we see in addition the circular segment of the two ventral columns, which by their dark color are sharply outlined against the remainder of the fibrillar substance. To the right and left of these two columns we perceive small masses of dotted substance; we merely call attention to them and shall not describe them. Figure 21 furnishes no noteworthy modifications of the preceding. We simply see a few cells of the periphery sending out their prolongations into the fibrillar substance. The point at which they thus penetrate it has already been indicated in Figure 15. In Figure 22 we have a section of several dorsal connective filaments; among others a lower root of the crural nerve is here seen to pass along the ventral surface of the fibrillar substance without penetrating into the crural lobule. Does there exist an upper root of the same nerve, which follows the upper surface of the dotted substance? We do not dare to decide the question. One thing is certain, and that is that if the nerve does exist it is accompanied along its path by a great number of widely ramified tracheæ, of which we see a drawing in _tr._ In the three figures which follow (23, 24, 25) the ventral column presents an interesting series of modifications. First of all, in Figure 23, it is surrounded by the lower ventral lobule, of which the two masses are in a lateral position, and whose commissures pass underneath the column. We see in the same Figure 23 the two lower roots of the crural nerve, advancing towards the column. In the 24th section the two roots have reached the column, and two other nerves cross the crural lobule; doubtless their destination is the lower dorsal connective filaments, but of this we have no clear indication. In the 24th section two other crural roots also enter the lower ventral lobule. This section is very favorable for the examination of the ascending fasciculus which we have already noticed in the longitudinal sections. It seems to us certain that this fasciculus terminates in the middle dorsal connective filament. Its origin is more uncertain. It seems to spring from the ventral column, or else to come from crural roots which, after having traversed the crural lobule, reascend towards the dorsal lobe of the ganglion, describing a curve exteriorily concave. It is possible that this ascending fasciculus has both these origins. The 26th and last section shows us the ventral column on a larger scale; the two columns being distinct from each other, though united at the lower extremity by a commissure. The _ensemble_ of the figure strikingly