ANTHOMEDUSAE ,i 



specimen is well preserved (in alcohol); it is 12 mm. high and wide, with about 12 tentacles in each 

 of the perradial bundles. The stomach has the very flattened shape mentioned in the description, and 

 it carries a great number of medusa buds, although this fact seems to have escaped Haeckcl's attention. 

 As Haeckel's original specimen from the Canary Islands has presumably disappeared, I designate this 

 specimen as the neotype. 



(2) The other locality mentioned by Haeckel as ' in der Nahe der Kap Verden ' is about 700 miles 

 north-west of these islands (24'' N, 33'^ W). From this locality we have one specimen collected by 

 Iversen in 1871. It is about 12 mm. high and wide and badly preserved, but it evidently belongs to 

 the same species. 



Thiel (1938a) has referred a specimen, taken in the middle of the Atlantic Ocean between the Cape 

 Verde Islands and the West-Indies (17° 53' N, 39° 19' W) to B. platygaster. This is the only record 

 of the species since it was described by Haeckel. Medusa buds were not observed by Thiel. 



It has generally been supposed that B. niobe Mayer was the only species of Bongainvillia which 

 propagated by budding. Vanhoff'en (1912 a) therefore referred a specimen taken in the South Atlantic 

 (about 24"" S, 21^ W) to B. niobe, but evidently from his description it belongs more properly to 

 B. platygaster. My previous identification of a specimen with medusa buds (taken by the Swedish 

 Antarctic Expedition off the Brazilian coast in 17"^ 32' S, 34° 55' W) as B. niobe must now be cor- 

 rected to B. platygaster, and some young stages recorded from the same region by Vannucci (19^1 a) 

 as B. carolinensis are also possibly the same species. 



Before I proceed to a description of the peculiar asexual propagation of the ' Discovery ' specimens, 

 some remarks on the general morphology of B. platygaster as compared with other species of the 

 genus are necessary. 



In B. platygaster the umbrella is about as high as wide, with a very thick jelly, somewhat flattened 

 above and with approximately vertical sides. It is not cubical as stated by Haeckel; in cross-section 

 it is almost circular with only a slight indication of flattening of the interradial sides. The specimens 

 taken ofi" the Brazilian coast are 4-1 1 mm. high and wide, the East-African specimens 5-7 mm. 

 Without exception, in all the specimens the stomach has the very flat shape which is characteristic 

 of this species. In one or two specimens the mouth tube protrudes slightly as a short and narrow 

 prismatic tube, while the stomach itself is quite flat ; in all the others the mouth is widely open at the 

 lower level of the stomach. 



Almost all the specimens are immature, with the stomach quadrangular in shape with straight or 

 slightly concave sides. Two specimens, a male and a female, have well-developed sexual products, 

 and their gonads are decidedly interradial. In the male (PI. Ill, fig. i) the stomach is slightly, in 

 the female somewhat more concave interradially, but there is no interradial interruption of the 

 gonads. Although the upper surface of the stomach is broad and flat, it is attached to the subumbrella 

 along the arms of a perradial cross. There is no indication of a gastric peduncle. 



The oral tentacles branch dichotomously immediately from their points of origin; there is no 

 unbranched basal portion or 'trunk'. In the larger specimens they branch 5-6 times. There are 7-9 

 marginal tentacles in each of the perradial bundles in specimens less than 7 mm. wide; 9-13 in the 

 larger specimens. The ocelli are crescent-shaped with the concavity turned outwards. 



I cannot see any morphological difference between the specimens from the two geographical areas 

 represented in the present collection. 



The species of Bougainvillia which might possibly be confounded with B. platygaster are 

 B. carolinensis, fulva, and niobe. 



B. carolinensis (McCrady 1857) occurs along the Atlantic coast of North America, from Woods 

 Hole to Florida, and recently it has been recorded, though with a query, from the Brazilian coast, in 



