ii6 



NA TURE 



[JSJov. 29, 18S3 



principal stem, the lateral buds of which never surpass it in 

 growth. Ill the same way the first formed hydranth of each 

 lateral branch retains its position at the tip of that branch, and 

 must be distinguished as a principal hydranth of secondary 

 order, becoming such so soon as it produces a hydranth bud 

 above its distal gonophore. This distinction is necessary not 

 only because the primary and lateral hydranths often differ in 

 size, but mainly from the most important fact that the principal 

 hydranths are sexually sterile ; only the lateral hydranths pro- 

 duce gonophores. No such distinction of principal hydranths 

 occurs amongst the Campanularidse and the Sertularida;. 



The above brief historical sketch and preliminary explanation 

 is extracted from the introductory part of the work. The special 

 part, which forms by far the greater portion of the whole, treats 

 separately of the details of the series of species investigated. 



Shy 2 -- 

 ent.u.eht—- 



Blsf 



Fig 3. — Tip of a stem oiEitden-iriuin riiceiiiosum{siCX\\z\, not d agrammatic), 

 with the principal hydr.-iiiths, Hhy, and ten lateral hydranths. Shy i-to: 

 Bist^ blastostyle, with female gonophores or ova ; Kzo, germinal zone in 

 wider sense, i.e. extent of the main stem and hydrocope containing egg- 

 cells. The letters ctit and ect indicate whether in the lateral hydro- 

 copes of the specimen ovicells were present in the ectoderm or enduderm. 

 or in both. 



The results with regard to two of tlie.'e form^, Cordylophora 

 lacustri; and Eiidemiriuiii, will be followed here, the former 

 being cho-en mainly because the account of it is illustrated by a 

 woodcut, which it is advantageous to reproduce. The structure 

 of Cordylophora lacustris is well known from F. E. Schulze's 

 •nost excellent mo^t excellent monograph. Weismann finds that 

 the regular branching of the stock in this species depends on its 

 following the law that "a principal or terminal hydranth of a 

 principal stem or lateral Ijranch produces no buds but those of 

 hydranth.s, never those of gonophores, and that only the 

 hydranths, and not the jjonophores, can produce buds." The 

 zone of gemination of the hydranths lies in the hydrocope, just 

 below the neck. In the female stocks the germinal cells do not 

 take their origin in the gonophores, but arise in the coenosarc in 



the ectoderm of the zone of gemmation of a principal hydranth 

 and in this well defined and restricted region only. 



The ovicells are certainly not preformed in the embryo or 

 larva, but are formed in the zone before the lateral hydranth bud 

 begins to appear out of ectoderm cells which differ in no respect 

 from other young ectoderm cells. The ovicells migrate in the 

 ectoderm from their place of origin to that where the bud of the 

 lateral hydranth has begun to form, and, passing into the lateral 

 hydrocope as it grows out, enter the gonophore as soon as it is 

 develo))cd, their entire course of travel lying in the ectoderm. 

 Kvery ovicell becomes an ovum, and enougli ovicells migrate in a 

 group into the lateral hydranth to fill several gonophores ; those 

 not destined for the first formed gonophore move onwards past if, 

 and a part of them pass later into the second gonophore when 

 this becomes formed between the first and the neck of the 

 lateral hydranth. This change of position of the ovicells must 

 be partly due to active movement, since the simple shifting due 

 to grow Ih could not push the cells past the first gonophore, and 

 long before the first gonophore is ripe these cells are found lying 

 beyond it, whereas beforehand they lay below it (see Fig. 4, «:). 



3. 4.— A principal hydranth, Hhy, ; 

 lophora : A'Z, actual germinal 2 

 former position of the germinal 

 migrating ovicells. 



I lateral hydranth, .'i/iy, of Cordy- 

 also zone of gemmation ; KZ\ 

 , Sph, female sporophore ; a/s, 



The niigration must take place very slowly and in a particular 

 direction, for the cells are never found scattered irregularly along 

 the whole stem, but always together in a small troop, and they 

 never make their way by accide.jt into a hydrocephalis. The same 

 process is repeated in the formation of the second, and, if ovi- 

 ce'ls enough be present, nf the third gonophore. A fresh swarm 

 of ovicells is never introduced from the main stem into a lateral 

 branch, and no new ovicells are developed in any later.al hydranth 

 until it ceases to become such by developing a hydranth bud 

 above its distal gonophore. It then becomes a principal hy- 

 dranth of secondary order, and acquires at once a germinal zone 

 beneath its neck, which supplies the gonophores developed on 

 its lateral hydranth buds with ova by migration, just as in the 

 case of the primary principal hydranth. It produces no further 

 gonophores it.self, and differs in no respect from the prim.iry 

 principal hydranth excepting in that it was once a lateral hy- 

 dranlh, and produced a set of gonophores, whilst the primary 

 principal hydranth never was lateral and never produced gono- 

 phores. The ova ripen in the ectoderm of the sporophores. 



The primitive male germinal cells in Cordylophora are formed 

 like the femile fr^ra young ectoderm cells, but their place of origin 



