No. 2.] LINE SYSTEM OF BATRACHUS TAU. 249 



In regard to the origin of the strand, my observations on 

 the embryos and larval forms of Batrachus would tend to con- 

 firm the opinion expressed by Leydig. Whatever the function 

 may be, its origin from the sense organ rudiment is not to be 

 doubted. In very young embryos the growth of the sensory 

 tissue is easily demonstrated, as shown elsewhere. In the 

 larval fish 15 mm. long, just after the yolk has become 

 absorbed, the strand is distinctly seen in preparation of the 

 skin, the cells of the strand between the organs still retaining 

 much the same appearance as in earlier stages (Fig. 15, con.st.). 

 Later, however, the cells undergo a change so that the tissue 

 appears as seen in Fig. 16, co?i.st. 



Comparison with Other Teleosts. 



1. Lophius piscatorius. — The goosefish resembles the toad- 

 fish in being destitute of scales and in having similar tentacu- 

 lar appendages in various parts of the body (Guitel, 30). The 

 sense organs are not enclosed in canals, but are protected 

 by projections of the skin, as in the case of the free organs 

 of Batrachus. The maxillary portion of the infraorbital line of 

 organs is greatly developed and the suborbital is wanting. 

 The innervation is quite similar in the two forms, the dorsal 

 branches of the VII being quite distinct from the V. 



2. Cottus gobio. — Bodenstein (9) has described the " con- 

 necting strand " in the adult Cottus and represents it in his 

 figures as on the floor of the canals. From his description of 

 the skin and the appearance of the canal organs, there is a 

 striking similarity between the two forms. 



3. Amiurus. — Batrachus and this common fresh-water form 

 have several characteristics in common. The naked skin, 

 closely studded with gigantic gland cells, the depressed head, 

 and general shape of the body is the same, but the sense organs 

 of the trunk in Amiurus are, for the most part, in canals. The 

 interesting comparison is in respect to the course of the 

 R. dorsalis recurrens facialis, which has been wrongly called 

 " trigemini." In Amiurus, according to Wright (31) and Pol- 

 lard (32), this arises from a " posterior dorsally placed gangli- 



