98 CEYLON MARINE BIOLOGICAL REPORTS. 



primary leaves like tho3e in otfa^er spp,cies of AQqfiia, and only develop phyllodes later on. ^t, tl^erefore, 

 seems that, also in Caulerpas, the h^e of the ghoot has retained most of the character of the pa^'cnt ^orm. 



The differeupcs in the formation of the shoots, which I have just dealt with, are characterized by 

 the different kinds of shoots originating successively in the same main axis (" allmahlich erfolgend," 

 Reisk:^) or, in other wofds, they are an example of differences in organogenesis at different stages of 

 develop;nent in th^ hfe of the slioot, ergo, analogous with the early forms (" Jugendformen "). 



Closely allied are the cases in which, from the same horizontal axis, vertical axes shoot out, eacli 

 with its several special kin<^ of branchlets (a " sprungweise erfolgende " variation, Reinkb) ; foy, usually, 

 the different vertical axes are to some extent individvialized, so that the branchlets in different axes a,ve 

 transformed at unequal rates. Thus, for instance, we can see one axis in which the primitive form of 

 the branchlets covers the a^is rathej;' high up, a,nother in which they only ocx^ur at the base and abruptly 

 pass over into the definitive ^i^m, (Cf. |lg. H of (7. Lesspnii, the branchlets to the left). 



If, now, the case occurs that in one axis the primitive forms are very few, or even, entirely \vantiipg, 

 and that the opposite is found ii;i another, that is to say, that there is nothing but primitive fornis, it 

 results that from one and the same horizontal axis quite different vertical axes may derive. In this form 

 of variation, however^ there are always transition forms, which show how the different axes may have 

 arisen. Th^ same is the case when the difference between the vertical axes only consists in their branch- 

 lets having developed in a dissimilarnumlper of rows (Cf • fig- 45 a, b, of C. longistipitatQ,). It is precisely 

 these differences that Rei;n]^e has kep,1? in view iii dist;lngui,sh;Lug the vj^riati,ons in the two kinds 

 " allmahhche " and " sprungxveise " (loc. cif,., p. 87), aijid he characterizes the difference thus, that in the 

 one we have to deal with a httle, in the other with a big bound (" sprung "). A.ccording to him, there 

 is no other difference. 



It is one very characteristic tra,it in all l;.he shoot variations we have been speaking of that they 

 occur successively, i.e., with transitions that may be more or less pronounced ; another that they may 

 to sorne extent be connected, with phanges in the mode of hfe, to \yhich they can be congeived as adap- 

 tations or '' ecologisms." The more primitive forms of branchlets at the base may be explained by the 

 base of the shoot having kept more of its primitive character, a character more closely aUied to that of the 

 ancestral form. Therefore they must be explained phylogenetically. 



(2) DiscQntinuous Variations — Atavistic reversions or mutatioris 1 



Quite different from all the other kinds of variations are thoi^e that occur in C. Lessonii /. 

 uticorinensis (fig. 12). Here we have three different kinds of branches, narrow ones with two-sided and 

 three-sided branchlets, like those in C. cupressoides , and furthermore, and this is the most noticeable 

 point, considerably broader ones whose very presence shows their affinity, with C. Lessonii. The 

 last-named kind occurs abruptly without any intermediary transition form, and corresponds in this 

 respect, as Reinke has pointed out ("Ueber Caulerpa," p. 85) in a striking way with the bud variations 

 we know in such higher plants — especially trees — as can be referred to under the name of " sporting 

 1^1 ants " (Darwin). 



Another example of the same phenomenon is the plant I have called C. sedoidea f. mixta (figs. 49, 50) . 

 The branciilets are, as a rule, ball-shaped with or without stalk, but among them shoots out occasionally 

 a perfectly cyUndrical branchlet, three or. four times as long as the others, and without any tmn^ition 

 at all. C. sedoides f. mixta thus shows an example of an abrupt variation of branchlets or pinnules, whereas 

 C Lessonii shows an example of similar variation but of long branches or whole assimilators. Thait 

 these variations are of quite a different kind to all the others I haye dealt with in the foregoing chaptei; 

 is evident, and the difference hes not only in the abrupt transition, but rather in that no correspondence 

 witli any external conditions at all , no adaptations, can be traced in this metamorphosis. 



Bud variations (Darwdj) or the ao-called vegetative mutations have been dealt with at length 

 in DE Vries' " Mutations-theorie " (Band, IX., p. 670). According to him, these can be divided iii,to three 



