2 6 



leads to a specially regular transverse segmentation of the stalk. These are (I) B. benedeni 

 Foettinger; which is discussed by Ritchie ; and (II) Vrnatella gracilis Leidy. In the latter case 

 in particular it is clearly shown from the descriptions of Leidy l ) and Dayexport -) that the 

 segmented condition is the result of constrictions formed in a stalk which is at first like that 

 of Pedicellina. In both B. benedeni, which constituted the type of Ehlers' genus Arthropodaria, 

 and l 'matcl/a, the stem-joints may give rise to buds which become new individuals, in much 

 the same way as in />'. bnlóosa, the genotype of Barenfsia. It is not impossible that the con- 

 dition seen in Pedicellinopsis is another development of the same process, if it be permissible 

 to regard the erect sterns which bear the Barentsia-Wke. individuals as modihed stalks of indi- 

 vidual units of the colony. Of this there is no direct evidence, so far as I am aware ; but the 

 account given by Busk 3 ) of the structure of the main "sterns" (see particularly figs 10 and 11 

 of his PI. IX) suggests that this may be their real meaning. P. fruticosa would, on this view, 

 be strictly comparable with B. bulbosa, but distinguished from it by the much more profuse 

 development of lateral buds from an original stem. 



Urnatclla, which is referred by Davenport ') to the Pedicellinidae, doubtless deserves 

 generic recognition, from the peculiar characters of its stolon, represented by a basal plate of 

 attachment ; as well as by reason of structural characters of the calyx. The question whether 

 Pedicellinopsis should be regarded as a valid genus appears to me a more open one ; since 

 the general architecture of the colony is very different from that of the forms usually referred 

 to Barentsia. I have accordingly not included it in the synonymy at the head of this account. 

 But whether P. frziticosa should be included or not in Barenfsia, this genus includes a con- 

 siderable range of modifications of the stalk, as follows : 



(I) stalk simple, without joints or lateral buds (usual form of B. gracilis M. Sars) ; 



(II) stalk unjointed, but producing lateral buds (B. bnlbosa Hincks; B. variarticitlata Andersson); 



(III) stalk jointed, bearing lateral buds (Gojiypodaria raniosa Robertson) ; 



(IV) stalk more regularly jointed, with few or no lateral buds {Arthropodaria benedeni Foettinger). 



These are probably to be understood merely as different conditions which ma}' be 

 assumed within the genus, and not as subgeneric groups ; since, as Foettinger r ') and Ritchie 

 have shown, the stalk of B. gracilis (assuming, with Ritchie, that this species includes B. 

 belgica Van Ben.) may be unsegmented or segmented; and since the difterences between my 

 third and fourth groups are mainly difterences of degree. 



It may not be out of place to enter a protest in this place at the use of the word 

 "polypide", by certain authors, to signify the calyx of a member of the Pedicellinidae. The 

 calyx consists of the body-wall of this part of the animal, including the alimentary canal and 

 nther organs. The term "Polypide" was introduced by Allmax c ), at the Meeting of the British 



i) Leidy, J., 1884, ^Urnatella gracilis"^ J. Acad. Nat. Sci. Philadelphia, (2) IX, p. 9. 



2) DAVENPORT, C. B., 1893, "On Urnatella gracilis", Buil. Mus. Comp. Zool. Harvard, XXIV, p. 19. 



3) Husk, G., 1886, p. 42. 



4) Davenport, C. B., t. cit., p, 30. 



5) Foettinger, A., 1886, "Sur 1'Anat. des Pédicellines de la cute d'Ostende", Arch. de Biol., VII, p. 301. 



6) Ali.man. G. J., 1851, Rep. 20"' Meeling Brit. Ass., p. 307; see also his u Mono^r. Fresh-Water Pol.'', 1S56, pp. 8, 41, 42. 



26 



