CARE OF THE BROOD iz7 



maining four species of the Zenometrinae and Bathymetrinae are known from single 

 specimens and each is a male. It may be that when the females are found some or all of 

 these species will be discovered to be viviparous. The five large species of the subfamily 

 Heliometrinae are not viviparous. 



Dr Mortensen has shown in a recent Discovery Report (1937) that there is in Ant- 

 arctic ophiuroids a similar very high percentage of viviparous forms. The Antarctic 

 ophiuroid fauna is much richer in species than the crinoid fauna. About 50 per cent 

 of them are viviparous. The highest percentage elsewhere is 15, in New Zealand. 



When Mortensen (1918, p. 2) made known the first three viviparous forms he re- 

 marked that it added to their interest that each had a separate way of caring for the 

 brood ; he was indeed fortunate, for he had three most interesting species before him. 

 The methods of the eleven viviparous forms now known are compared below. 



In only one group, the species of Isometra, does the viviparous habit appear to have 

 led to a modification of the hard parts. The segments of the genital pinnules of the 

 females are strongly expanded to arch over and protect the ovary and brood-pouch. 

 There is a much smaller expansion of the segments carrying the testes in the male — 

 so much smaller that whereas in almost all comatulids males and females are indis- 

 tinguishable to the naked eye, in the species of Isometra they may be recognized at a 

 glance. 



In all the viviparous species there is in the female a brood-pouch beside each ovary. 



In Notocrinus virilis and N. mortensefii the gonads have a unique position. They do 

 not lie, as in all other crinoids, along the pinnules, but in the axils between the pinnules 

 and the arms, largely on the arms. The brood-pouches of the females are distal to the 

 ovaries. The two species form an interesting contrast in the extent to which they protect 

 the brood. A brood-pouch of a big A^. mortenseni may contain as many as ninety-two 

 embryos in all stages of development and varying in size from 0-25 to 0-48 mm. The 

 largest have five broad bands of cilia ; they presumably go through a free-swimming 

 stage before metamorphosing into pentacrinoids. The brood-pouches of A^. virilis each 

 contain only one to three, usually two, embryos, all at the same stage of development. 

 They are up to 2 mm, in length, that is, four times as long as those of A^. mortenseni, 

 and have no trace of ciliated bands. It must be supposed that they drop out of the brood- 

 pouches to the sea floor and change into pentacrinoids there. A series of pentacrinoids 

 is described on pp. 210-219. 



In the other viviparous crinoids the brood-pouches lie alongside, or beyond, the 

 ovaries on the pinnules. They lie alongside the ovaries and always on their distal sides, 

 i.e. on the side of the pinnule nearest to the arm from which it springs, in Eumorpho- 

 metra concinna, Phrixometra longipinna, P. longipinna var. antarctica and the four species 

 of Isometra. 



In these species each brood-pouch contains a fair number of small embryos which 

 possess ciliated bands. (I can see no ciliated bands in those of Phrixometra longipinna, 

 but I think that is because they are too young.) It is probable that in these species, as 

 in Notocrinus mortenseni, the embryos pass some time in a free-swimming stage before 



