TIIK SEKTULARID^E. 43 



the hydrothecse are arranged incompact groups of pairs, a group to cadi internode, the upper 

 pair of a group being noticeably smaller and ditl'crcnt in shape from the lower. This vcrv 

 characteristic form is the basis of the ^enu> /'/.*/////,/ of Lamouroux. 1 



A large niimhcr of si>ccics agree in having the operculum of three or four (laps and the 

 hydrotheea' strictly alternate. These form the well marked and well known genus \< /////,//, ///.- 



Again we tind that a comliination of the characters of the operculum together with the general 

 form of the hydrothecse can he used to advantage. A large numlierof specie- agree in having an 

 operculum composed of a single Hap which is hinged to the adcanlinc side of the margin. These 

 have all been included l>v Levinsen in tin' "'enus Divhasia. A studv of these forms leads to the 



."> f . 



discovery that the genus thus constituted is made up of two well-marked groups which agree in 

 the character of the operculum just given, Imt ditl'er widely in the form of the hydrothec;e; one 

 group consisting of species with tubular hydrothecie that are not distinctly swollen below, and 

 that have very large apertures without any distinct neck or constriction of the distal part of the 

 hydrothecal wall. These forms can very well be retained in the genus J>!j>/i/isi,i, substantially as 

 originally described by Agassi/. The other group having an operculum of one adcauline valve 

 consists of species which differ from I >iji/mxi/i, and agree among themselves in having hydrothecse 

 that are more or less bottle-shaped that is, their lower portion is bulged out or swollen like the 

 body of a flask, and their distal parts are constricted so that the diameter of the aperture is much 

 smaller than the diameter of the lower portion. There is often also a distinct constriction can-ed 

 by the thickening of the hydrothecal walls below the margin. The forms just described constitute 

 what seems to me to be a very well circumscribed genus .!/(/,////<///,/.' 



Si.rtli. As a last resort we lind that a combination of the position of the hydrothecse, 

 whether opposite or alternate, and the character of the internocles will serve to aid in solving 

 the last and most perplexing problem of all, the separation into generic groups of the forms 

 still remaining, which agree in having the hydrotheea' in two rows, margins smooth or dentate, 

 the operculum of one adcauline Hap, or of two (very rarely three) flaps.* We find upon exami- 

 nation of the very large number of species agreeing in these particulars a number that agree in 

 having strictly opposite hydrotheea 1 that are not greatly embedded in the hydrocaiilus, an 

 operculum comprised almost always of two valves, and the internodes normally and commonly 

 bearing two hydrotheea 1 ; or the hydrothecie may be slightly snbopposite. but the internode- 

 are regular and normally and commonly bear two hydrotheea'. showing that the latter are 

 essentially paired, as pointed out by Hale."' Such forms can be referred to the original genus of 

 the family, !<< rlnliir'm. 



The remaining forms agree in having the hydrotheea' normally and regularly subopposite 

 to alternate, usually more or less embedded in the hydrocaulus; operculum of one flap (abcau- 

 line) or of two flaps; more than two (often many) hydrotheea' to each internode. the latter 

 being of very irregular length even in the same colony. These forms we can refer to the old 

 genus Tlutiiir'm. 



It will be noticed that seven of the nineteen genera admitted by Marktanner-Turneretscher 

 (see p. ;-!',) are not included in the scheme as outlined above. These seem to me to be untenable, 

 and the species included in them can be disposed of as follows: 



(irnnuiiiiriii and //'//'"/'/A''''* do not belong to the Sertulariihe. 



< 'ith/jifnf liiiiiirin and Si/m/il, <-fn.-'i//i/<iis should be united with X rl<il<n-< /!</. from which they 

 are not separated by what appear to me to be adequate characters. 



MIIIIIIJMIIIHI is quite a typical /'////A//'/,/, according to the scheme here adopted. 



f),-Niiiiixi-;/j>lnin can with propriety be referred hack to the old genus X, // /,///./. 



1 Nouveau bulletin philomatique, ili'ccuilnv, ISIL", p. is:;. 



-< iv:iy, Itadiata, I.Ut nf s|M-ciiiii'iis "f I'.ritish aninmls, etc., l-mnlim, ls.47, p. us. 



:1 Kirrhenpiuirr, Nonlisrln- ( oHuiiLrrn un.l Artcn vim Srrtiilariilcii, I laiiilmrt;, ISsl. p. L'!i. 



'The author does not claim to bare successfully solved this pmlilimi. \\liirli appr:u> \\itli "iir present kimwl. 



edge to be insoluble, but hopes that the arrangement suggested will l>r practical in fact, altlnm^li unsatisfartury 



in thenry. At anv rate, it is the lnvt that he has licen ahle tn ilevise after very careful punilerin}.' nf the suliject. 

 ''Australian Ilyilniiil /nnpliytes, lsS-1, pp. 115,11(5. 



