FRESHWATER AUi.K. I!) 



material were however considerably larger than those figured by .Messrs. West (<>/>. cit., 

 PL XXIV r ., fig. 10). We may add to the differences between the two forms the 

 habitat on moist stones, which has already been commented upon by J. G. Agardli, 

 De Toui and Messrs. West. 



The greatest difficulty in the way of regarding P. <iiil<nrtic<i as a Aperies 

 distinct from P. cr/.spa is the absence of all indications of a separate course of 

 reproduction in the former. Abundant reproduction by the characteristic packets of 

 cells was going on in the material from Mt. Terror, but, except that the production 

 of the packets in this case was certainly confined to the margins of the thallus 

 (text-figure F), there was nothing to distinguish them from those of P. cri*]><i. For 

 this reason I think it may ultimately appear that P. ant<tr<-ti,-a is only a varietv of 

 P. crispa, although one cannot agree with Wille (up. n't., p. 218) that the two 

 forms are identical. 



13. PRASIOLA CALOPHYLLA. 



(Text-figures G-J.) 



Piasiola caloplujlla (Curm.) Mcii^b. ; Rubcnhorst, op. cit., pp. 310-311 ; Du Toni, Sylloge AliMnim, 

 i. (1889), p. 144. 



//'i/>. Attached to rocks, freshwater pond in gap, Winter Harbour, January 

 12th, 1904; attached to stones, Granite Harbour, January 20th, 1902. 



This form was markedly distinct from the two previous ones in the thread- 

 like form of the thalli and in being firmly attached to the rocks by a long 

 filiform stipe ; this slowly widens (consisting at first of one, then of two and then 

 of four longitudinal rows of cells) into a fairly broad thallus (25 or more rows 

 of cells wide), which is in no way sharply demarcated from the stalk. A number 

 of thalli arose from the same point, appearing to the naked eye like a bunch of 

 coarse filaments (text-figure G). 



The base of the stalk is slightly enlarged to form an organ of attachment. 

 The cells of the stalk are prominently flattened and have thick walls (diam. 

 stalk = 20 /x; diam. cell-contents = 9-1 1 /t ; long, cell-contents = 3-5 p), but the 

 longitudinal walls are thicker than the transverse ones (text-figure H). Here 

 and there longitudinal division takes place in the cells of the stalk, so that 

 two rows of cells arc present, but such divisions are merely local (leading to 

 slight bulgings of the stalk) and generally only extend to a few cells, bevmid 

 which the stalk again consists of a single row. It is only at a considerable 

 distance from the point of attachment that the stalk becomes permanently biscriate 

 (text-figure I), and this condition may again persist for some distance before 

 further longitudinal divisions take place and a quadriseriate arrangement is attained. 

 After this, longitudinal walls appear to be formed more frequently and the stipe 

 rather rapidly broadens out into the fiat thallus. Even after the bi- or quadri- 

 seriate condition is arrived at, there may often be local reversions to a nni- or 



