7. Anthozoa (incl. Hydrocorallia). B. Alcyonaria. 37 



genus and are available for generic diagnosis. Other characters utilised are 

 the mode of branching, shape of calyces, axis, etc. A key is given to the 

 genera of this family and the following are described Gorgonella 4 (2 n.), 

 Verrucella 3 (1 n., 1 n. nom.), Gtenocella 1, Juncella (redefined) 4 (1 n.), Seirpearellak 

 (2 n.), Nieella 2 n., Ellisella 1 n., Plumigorgia n. 1, Isidoides n. 1 (closely re- 

 sembles certain Isidae, but the axis is unjointed; also shows an approach to the 

 Primnoidae in the character of the operculum). 1. Scirpearia is regarded as 

 untenable, some of its species can be placed in June., others in Scirpearella. 

 All but one of the 21 species secured by the Siboga are found at a depth less 

 than 100 m. Scirp. gracilis has a great bathymetric range 34 to 1886 m. 

 The apparent centre of distribution for this family is in the Oriental region. 

 Simpson( 1 ) states that the Juncellids are typically shallow water forms, oc- 

 curring chiefly in tropical, but also in temperate, seas, generally within the 

 100-fathom line. In the Mergui Archipelago colonies are left uncovered, at 

 low tide, for two hours per day, but survive. Colonies may be over 6 feet long; 

 they are very flexible and possess great powers of regeneration. In the Mergui 

 they often bear, as commensals, Cirripedes and the Lamellibranch Pteria macroptera, 

 the former being almost overgrown by the ccenenchyme. The colonies examined 

 were dioecious and not viviparous. The nature of the surface, consistency and 

 colour of the coenenchyme are of no taxonomic value but the canal system 

 is of great importance. It consists of (1) an inner longitudinal series of canals 

 separating the inner canal-bearing part of the cosnenchyme from the axis, (2) an 

 outer longitudinal series separating the two divisions of the coenenchyme, (3) a 

 transverse series ramifying in all directions through the inner part of the 

 coanenchyme and uniting (1) and (2). The outer series of canals communi- 

 cates directly with the polyps and, by means of the transverse canals, also 

 with the inner series. The number of canals in the outer and inner series is 

 not proportional. The inner canals are not of equal value, some are much 

 larger than others, no matter at what level examined, and the number of large 

 ones 2, 3 or 4, - - is constant for the specimen. These large canals are 

 symmetrically arranged and corresponding to them are the longitudinal tracts 

 devoid of polyps. The furrows in the axis correspond to the inner canals, 

 and as the latter diminish in number from the base upwards so also does the 

 number of ridges and furrows and is therefore not a reliable character. Each 

 polyp, which is very retractile, is divisible into a lower cup-like verruca 

 and an upper tentacle-bearing anthocodia. The polyps are always arranged in 

 a certain number of longitudinal series, definite for the species and depending 

 on the number of main longitudinal canals. A spiral arrangement, when pre- 

 sent, is secondary and due either to overcrowding of the polyps or to torsion 

 of the colony. The flexible axis consists of a horny substance impregnated 

 with calcium carbonate. The increased thickness of the older part of a colony 

 is due almost entirely to increased thickness of the axis, not of the co?nen- 

 chyme. The spicules are extremely small; the different types - - clubs, double 

 clubs, double wheels, spindles - - retain their individuality no matter from what 

 level of the colony they may be taken, those from different levels of the same 

 colony, or from colonies of different ages, show marked deviations in absolute 

 size but not in proportionate size. For specific determination the prime 

 factor is the character of the spicules, next their average size, while the pro- 

 portionate number of each type is unimportant as it depends on the proportion 

 of ccenenchyme and verrucse taken for the preparation. The affinities of the 

 Gorgonellidse seem to be with the Pseudaxonia. The Juncellids are divided 

 thus (1) Spicules include clubs, Juncella. (2) Spicules do not include clubs 



