^o8 THE ALGAE 



evidence in favour of scheme A. Scheme B in particular involves 

 the development of parenchymatous types (Ulvales), their sub- 

 sequent reduction to a simple filament (Ulothriceous type) and the 

 re-elaboration of this to a branched, partially coenocytic thallus 

 (Siphonocladales). Whilst not impossible, such a sequence seems 

 an unnecessary postulate. In scheme A it will be observed that 

 several lines of evolution divaricate from the primitive source. One 

 line leads to the development of aggregate spherical colonies, cul- 

 minating in Volvox: another Une leads to the production of gelatin- 

 ous colonies, which may become net-like {Schizodictyon\ and an- 

 other to dendroid colonies {Prasinocladus). All these, however, 

 would seem to represent *bUnd alleys' in evolution. In the main 

 sequence of evolution the logical development is the non-motile 

 unicell and then the attached unicell. From the former have de- 

 veloped more colonial forms terminating in giant nets {Hydrodic- 

 tyon\ but simple siphonaceous forms may also have been produced 

 (Protosiphon). The non-motile unicell would easily develop into a 

 simple filament {Ulothrix) from which a number of specialized 

 'blind' end fines must have appeared (Oedogoniales, Conjugales). 

 From this central key position evolution would appear to have pro- 

 gressed in three directions : 



{a) To parenchymatous types, e.g. Ulva, Enteromorpha. 

 lb) To partially coenocytic types, e.g. Siphonocladales, and 

 thence to the Siphonales. It has been suggested as an alterna- 

 tive that the Siphonales may have arisen from the siphon- 

 aceous Chlorococcales {Protosiphon). The principal argu- 

 ments against this hypothesis would seem to be: 

 (i) there are no intermediates between the wholly haploid 

 Protosiphon and the wholly diploid Siphonales. One 

 would expect to find some forms with both generations 

 alternating; 

 (ii) the Chlorococcales are nearly all fresh water whereas the 



Siphonales are all marine; 

 (iii) intermediates can be envisaged from at least some mem- 

 bers of the Siphonocladales and the Siphonales (scheme 



More detailed discussion of this problem wiU be found m 

 Chapman (1954). 

 (c) To branched heterotrichous types (Chaetophorales) with 

 subsequent reduction of one or other of the generations. 



