49 



shown the existenc-e of an eye (op. cit. p. 314, Taf. X. fig. 21 and 23, oe), which in the illu- 

 stration is drawn as situated on the ventral side of the forehead »in Form von zweien am 

 oberen Theile etwas verdickten sichelforniigen Pigmentflecken , welche in der Mitte sich 

 beriihren iind eine x-fcirmige Figur darstellen«. As a matter of course, the e3^e must be 

 found on the dorsal surface, but I have been unable to find it on my larvae, pi-obably 

 because the spirit had dissolved the pigment. 



The Antenmdcc are always rather short; they consist typically of three joints, the 

 second of which is usually short and not unfrequently coalescent with the first, in which 

 case we only perceive two joints. The first, and particularly the third joint, are provided 

 with pretty long sets ; the terminal seta of the third joint is very long, and from the lower 

 side of this joint proceeds always a single, particularly long olfactory seta (pi. I, fig. 1 1, b), 

 wliich is at least double, usually several times, the length of the whole antennula; sometimes 

 this seta is exceedingly long, as e.g. in Sphrer. dispar (pi. IX, fig. 3 k) and va. Splwer. insignis 

 (pi. X, fig. le), where it I'eaches fiuther than the middle of the abdomen, nay in the last- 

 mentioned species the olfactory seta in itself is longer than the whole cephalothorax. The 

 antennulae are always attached pretty far from each other at the edge of the cephalothorax, 

 the area between them forming what I call the fi'ont. Close to the inner margin of the 

 base of the antennula we often see an oblique list; moreover, in nearly all species of 

 SplifpfoneUa which are parasitic on Cumacea, the front is decorated with one or several 

 rows of delicate and peculiar processes, wliich decoration reaches its highest development in 

 Sphter. cJecorata (pi. VIII, fig. 3i and fig. 31). In Splicer, modesta these processes are replaced 

 by transverse lists (pi. IX, fig. 2i). 



The AntP)ui(P proceed beliind, and usually at the same time somewhat oblifiuely 

 inside the antennulie, but never from the base of the rostrum. Sometimes they are conside- 

 rably shorter than the antennula3, sometimes about the same length, and in the genus 

 Homoeoscelis (pi. II, fig. 11 and pi. XIII, fig. Ih) more than double the length. In Splicer, 

 marginaia (pi. XIII, fig. (ig) only two joints are found, in all other species they consist of 

 three or four joints, three of which are always distinct, but it is often difficult to make out 

 for certain, whether the eminence from wliich the supposed second joint proceeds, is a real 

 joint, or in other words, if the apparent basal joint consists of two joints; as, however, 

 Spheer. antillensw has four very distinct joints (pi. Ill, fig. 2e), this is probably the typical 

 number. The terminal joint is nearly always short and usually ends in a long, thick seta, 

 beside which we frequently find a shorter one. The next joint is now rather short, now 

 long, or very long, and where the antenna is long, it is on account of the length of tliis 

 joint, as the basal joint, or where there are four joints, the two first of these are never 

 elongate, but sometimes (pi. IX, fig. 2i and fig. 3k) comparatively broad. In Mysidion 

 (pi. XII, fig. 2h) the antennae are very small, and in Sphter. mkrocephala (pi. VIII, fig. 2i) 

 almost rudimentary, in both cases 3-jointed, -nnth an exceedingly short terminal seta. 



The nostrum seems to correspond only with the more distal pai-t of this organ in 



7 



