The development of 16 spp. of Molgula and 2 of Eugyra is described. Ten are viviparous. Viviparity is due to reduction in adult size (in this family), in turn due to precocious sexual maturity resulting from the colonization of exposed rocks by sand-adapted types. Increase in egg-size is a result of viviparity. Nine species develop without forming tailed tadpoles. Two species said to be anural were found to produce tadpoles. Cleavage is identical in anural and urodele development. Hatching in anural development is always by inembrane rupture and never by digestion. An aggregation of cells is always seen near the larval stomach in both types of development; in the urodele it is known to be the absorbed tail. Failure of tail development is due to the absence of swelling on the part of the notochord cells, which, however, are present in normal number. Failure to extend on the part of the notochord results in a delay in closure of the blastopore. Imbibition by notochord eells is to some extent dependent upon the alkalinity or acidity of the surrounding medium. Out of 139 spp. of the Ascidiacea anural development is confined to 9 molgulids. Within the Molgulidae the origin of anural development is poly-phyletic. The otolith (the only sense organ) is present in all urodele. larvae, absent in all anural larvae. Anural development is to be correlated with large size, oviparity, and a free, unattached sand life; urodele more with small, viviparous, attached forms. The first type is considered to be primitive, and to have secondarily colonized the rocks. It is concluded that anural development is a direct response to the sand-embedded habitat, that the hatching enzyme is lost when tail development is disturbed, and that anural development is compatible with race-survival only when there is an efficient alternative method of hatching.