1885.] PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 135 



The followiDg observations of bis owu bare been put upon record in 

 the latter's report, cited below ; but it is evident that be overlooked 

 the pecubar vascular lobes appended to the vertical fins of tbe embryos 

 noticed by Blake, and also drew some erroneous inferences as to the 

 identity of the yelk-sac, which will be noticed later. From pages 165 

 to IGG, inclusive, of Girard's report the following i^aragraphs are 

 quoted : 



"But the most remarkable trait in the organization of the fishes of 

 this family consist in the mode of reproduction. The eggs, instead of 

 being laid, as is the case in most fishes, are retained within the body of 

 the female, where they undergo not only their embryonic growth, but 

 likewise a growth which might be termed larval, it being subsequent 

 to their escaping from the egg envelope, until they have attained a 

 size sometimes of several inches in total length. Nevertheless this 

 peculiarity of habit is not altogether without analogy in the class of 

 fishes. Moreover it bears no resemblance whatever to the marsupia- 

 lian gestation; in the first place the eggs develop in the ovary, not in 

 a uterus, and there is no placental connection of any sort; secondly, 

 the young are not at liberty to quit and enter the ovarian cavities alter- 

 nately and at i^leasure. 



" The observations which we have made upon the genital apparatus 

 of the female have satisfied us that there exists an ovarian sheath or 

 sac, whicli during the early period of pregnancy is an elongated and 

 subcylindrical tube, containing the ovaries proper, two in number, 

 each of which consist of two, three, or more vascular membranes, at- 

 tached by their upper edges to the upper floor or roof of the sheath, 

 forming either one or two pouches (according to the number of these 

 membranes) of the same length as the sheath itself, widely open be- 

 neath, though not in direct communication with one another, since the 

 membranes hang loosely down, reaching the lower floor of the sheath. 



" The eggs are formed within the texture of the ovarian membranes 

 themselves. We have examined the ovaries of Unnichthys heermanni 

 and Einbiotoca argyrosoma when the sheath within which they were 

 contained was not larger than an ordinary quill. Numerous eggs could 

 be observed in a very immature state, appearing to the unarmed eye 

 like minute dots. Under the microscope tbey exhibited evident traces 

 of the germinal vesicle, surrounded as yet with a very scanty supply of 

 vitelline substance. 



" The sheath and the ovaries are gradually increasing in bulk, as 

 the eggs themselves first increase in size and the embryos afterwards. 

 The sheath is chiefly a muscular membrane, while the ovaries, we have 

 stated, are altogether vascular. 



" When mature the eggs either fall into the space between the mem- 

 branes or ovarian pouches, or else remain attached to the ovaries until 

 the embryos issue out of them. We are inclined to think that they 

 dro]) into the pouches as eggs. At any rate we found very young em- 



