284 DISCOVERY REPORTS 



the area of distribution in summer and autumn (Stephensen, 1924, p. 119). The young 

 up to and including the 4 th insta r remain within the house (Mogk, 1927, p. 130). 



These observations are borne out by the present collection. Females with ova were 

 found in October (33°-34° S) and November (3 S), with young in the 1st instar in 

 November (33°-34° S). Young in the 2nd instar were found in June (40 S), in the 3rd 

 and 4th instars in both June (35°-46° S) and October (13 N), in the 5th instar in 

 June (32 S) (latitudes approximate). 



Dudich (1926, p. 130) does not admit Stephensen's views, and states that in the 

 Mediterranean there are two "swarms". The autumn swarm rises to the surface and 

 propagates in November to December. The young sink in January and grow in the 

 deeper layers until October, when they rise again. The spring swarm rises in February, 

 propagates in March to April ; the young sink and remain below until February. Thus 

 at the surface from October to May there are two maxima or swarms, while in the 

 deeper layers from May to October there are all stages of both swarms, those of the 

 autumn swarm being more advanced. The times of appearance and bathymetrical 

 distribution are similar to those of Pyrosoma (p. 133). 



The 1st instar is passed within the brood pouch of the mother (cf. no. 18); the 2nd 

 to 4th instars are attached to the wall of the house. Stephensen (1924, p. 1 19) says that 

 ^the young are attached in two groups and refers to a figure in Minkiewicz (1909, fig. 22 

 on p. 2). Minkiewicz does not specifically state the position of the larvae, and his figure 

 may be misleading as it may represent a schematic sagittal section of the animal in its 

 house, after removal of one side of the house on which there may have been a further 

 batch of larvae connecting the two batches drawn in the figure. Stephensen's observa- 

 tions, if they are his own first-hand observations, are in conflict with what I find to be 

 the case in the Discovery material. Here there are twelve houses containing young, 

 and in every case there is only one group. The young are arranged in a radiating manner, 

 their heads pointing_in a centrifugal direction. Dudich (pp. 123, 124) says that the 

 larvae feed o n the substance of the house, and as they seek for new pastures the 

 "Brutgurtel" spreads and resolves itself into two girdles. 



In two cases, no. 3 (October) and no. 22 (June), the group of larvae is composed of 

 individuals in two stages of growth, namely the 3rd and 4th instars. Where the group is 

 composed of individuals all of the same instar, it forms a compact more or less circular 

 patch ; but where two instars are present the number of individuals is much greater and 

 they spread out over the whole inside of the house, though they are just as closely packed 

 as in the groups containing only one instar. 



The question arises where two instars are present, whether they all belongjo_cme 

 br ood or to two -broods. Appearances certainly seem to suggest two broods, but one 

 may doubt the possibility of a single female producing two broods in such rapid suc- 

 cession. Further observations on more abundant material might lead to a definite con- 

 clusion. Dudich (p. 124) states that the eggs are laid in three batches at three-day 

 intervals. 



Distribution. Mediterranean ; North and South Atlantic, 6o° N-36 S ; Indo-Pacific. 



