papillosa, U. suópapillata, U. explanata, Codium divari 



i var. spumosa. Hie other novelties described here, 



Chlorodesmis Hildebrandtii, Avrainvillea clavatiramea, 



I f :.'.:, A. pacifica, .1. laceraia var. robustior, .1. tumult- lp lux 



ii' iv. gen.) pe/tata, Rhipüia tomentosa f. tsonata, 



tespilosa, Tydemania Gardineri, Udotea indüa, U. 



identalisy I . verticillosa tuil description), U. Wilsoni. 



< nus a^ Penüillus, it became necessary to take into account 



long time included in Penicillus and seemed almost to link 



through the species A'. oblongus. Rhipocephalus is a purely West Indian 



never been recorded from tin- Eastern hemisphere; nevertheless we consider its 



in the present paper t<> be indispensable as completing tin- comparative study and 



: the flabelliferous members of the Udoteae, especially in relation to tin- Siboga- 



mania. 



We have been compelled to separate off from Udotea the Australian species U.peltata 



|. Ag., which in genera] structure really comes nearer t<» Rhipilia or Avrainvillea than to 



;n all these genera it is however distinguished by the possession of one remarkable 



liarity, which necessitates tin- creation of a new genus for its reception. To this we have 



given the name of R/tipiliopsis, and we have appended an account of this little known plant 



with figures <<ï its peculiar interna! structure. 



Two other s|)eci<s which wc have found reason to remove from Udotea on the ground 

 of their evident affinity with Chlor odesmis comosa, are the ancient and well-known i '. Desfontainii 

 of th< Medtterranean and the primitively developed and recentl\- discovered U. minima of 



lo receive these two species we restore the genus Flaöellaria^ instituted by Lamoui 



in 1813 for F. Desfontainii. F. minima is dimorphic and is particularK' interesting as retaining 



rms indisputable traces of its evident derivation from a simple CAlorodesmis-like 



-t"r. As wc show elsewhere (p. 48) these two green and uncalcified species are quite out 



of harmony with the true species of Udotea, which are calcified and form an exceedingly 



ng and connected series of developmental stages easily traceable back to the simple 



flabellate species U. javensis [Rhipidosiphon of Montagne). As now arranged, Udotea 



i tropical genus, whereas Flabellaria is contined to the Mediterranean and to the 



esl Africa. 



Turning now to tin- question of relationship between the various genera composing the 



1 t hink that it is natura! to separate them into three distinct groups, the 



eae and the Udoteae. The Codieae <-onsisi of two genera Codium and 



which are characterised by the possession of a cortica! layer of utricles, loose 



into a pseudo-parenchyma. Codium is a spongy feltwork which assumes various 



the various species, and its cortica! layer consists of more or less 



>ry utricles. contiguous luit not welded together into a continuous tissue. These 



ze according to the species. Growth is not confined to any particular 



docodium, where the principa! growth is at the apices and is continuous. 



