Kellia suborbicularis (Montagu) has long been known to be viviparous. Lovén (1848) mentions this fact and Jeffreys (1863, p. 227) quotes Mr Clark who “found in the ovary of one individual ova in an advanced state with fully formed testaceous fry.” It is probable, however, that these shelled larvae were in reality contained in the gill pouch where later workers have found them. Pelseneer (1935, p. 517) says that the eggs are incubated in the internal gill lamella, and this was also observed by myself.

It is striking that although the shelled larva is extremely small when liberated (shell 0.064–0.08 mm. in length), the late planktonic larva is very large (shell 0.30–0.37 mm. in length) and one of the most conspicuous bivalve veligers in the plankton. This indicates a prolonged free-swimming life quite unlike that of the oyster, Ostraea edulis, whose shelled larvae are liberated only to lose the velum in a few days. These late larvae of Kellia have been kept in bowls and plunger jars in the Plymouth Laboratory until they metamorphosed and for months afterwards, growing into beautifully clear and transparent individuals which undoubtedly belonged to this species.

Kellia suborbicularis is placed by Winckworth (1932) in the Erycinidae, but this family is closely related to the Leptonidae, Montacutidae and Galeommatidae, most or all of whose members are viviparous. Lovén (1848) described the young shelled stage of Montacuta ferruginosa and M. bidentata, the latter species now being referred by Winckworth toMysella. Both keep their young in the gill pouch until the shelled stage and Lovén's beautiful figures show that the early shelled stage in each of them is very like that of Kellia.