493 



that have been investigated but I have been unable to seize upon any 

 characters of definite value in classification. 



The orientation of the gonad plays a very small part in the di- 

 vergence of the species. When the gonads are numerous and small, as 

 in Pandocia and Paratona, they are for the most part oriented with their 

 apertures toward the atrial siphon. When they are large and elongated, 

 their apertures open usually near the base of the atrial siphon. This is 

 of advantage for the ready escape of the eggs and spermatozoa. A strong 

 current of water will usually be passing over the apertures. In some 

 forms, however, the gonads open some distance from the atrial siphon. 

 In Dendrodoa carnea the single gonad opens posteriorly and ventrally. 

 This results in the retention of the eggs, a broad cavity being formed 

 behind the pharynx. In Katatropa the gonads (two on each side) open 

 anteriorly and ventrally (Fig. X). As in the other genus, the eggs are 

 retained, although there is no definite brood cavity. In the related ge- 

 nus, Styela, the eggs are not retained. The difference shown in Fig. X 

 is as exhibited by the stalked species. In the shorter species the dif- 

 ference is more marked. 



Atrial Tentacles. Up to the present the atrial tentacles have 

 not been used in classification. The atrial velum shows constant dif- 

 ferences in the species, but as these differences are quantitative (nar- 

 rower or wider) they cannot readily be used. The forms with the testes 

 growing away from the ovary have a very narrow velum. In the others 

 it varies in width. The tentacles are sometimes restricted to the at- 

 tached margin of the velum (Fig. XI) and sometimes scattered over the 

 whole inner surface (Fig. XII). In Dendrodoa, Pandocia and Cnemido- 

 carpa the velum is variable in width and the atrial tentacles are in a 

 single row. In Paratona elata the velum is extremely broad and the 

 tentacles sparsely scattered over its whole surface. In Styela, Kata- 

 tropa, Pelonaia, Ooniocarpa and Botryorchis the velum is quite narrow 

 and the tentacles more or less thickly scattered over its whole surface 

 (Fig. XII, 1). 



Anus. Dr. Hartmeyer has suggested that the characters pre- 

 sented by the margin of the anus (whether lobed or smooth) is of some 

 value in classifying the Styelids. I have not found it so with my ma- 

 terial. The lobing of the anus varies from very distinct to indistinct. 

 I have found it even in Dendrodoa (D. carnea), in which genus Dr. Hart- 

 meyer states that the margin of the anus is smooth. Being a question 

 of degree only, it will not be readily applicable but will still be of as- 

 sistance in checking over a classification based upon other characters. 



Siphonal Spinules. These small spines are developed to a much 

 greater extent in the Tethyidae (Cynthiidae) than in this family. They 



