APOGAMOUS FERNS. EVOLUTION OF THE SEPARATE SPECIES 

 obtained from Italy and from Corsica. In my experience these numbers are very nearly 

 correct (see Fig. 174), and the wild European species may thus be regarded as diploid 

 in contrast to 'var. major' which is tetraploid. All other commercial strains that 

 I have investigated, notably var. Wimsetti and var. Childsii, are also tetraploids which 

 is the reason for thinking that the wild type had fallen out of use. 



The principal material of Pteris cretica used by me has been the following: 

 (i) The horticultural strains 'var. major' and 'var. Wimsetti' grown for the market, 

 together with Cyrtomium falcatum, by Messrs Clibran of Altrincham, to whom I am 

 extremely grateful for unrestricted access to their greenhouses. 



P. crehca n ^ 58 



Fig. 174. Meiosis in diploid Pfem cr^^ica Fig. 175. Explanatory diagram to 



L. permanent acetocarmine, from an Fig. 174. x 1500. . 



eight-celled sporangium to show de- 

 tailed chromosome count ('h' = 58). 

 X 1000. 



(2) var. albolineata from Kew and Ceylon together with var. albolineata-cristata, a 

 crested form of albolineata otherwise closely resembhng it, from Kew. 



(3) Wild European material collected by myself from a locality near the western 

 shore of Lake Maggiore in 1937 and since maintained in cultivation. 



(4) Wild tropical material brought alive to Kew from Uganda in 1938 but un- 

 fortunately lost as a result of the war. 



Although the horticultural varieties first mentioned (i) were the earliest to be in- 

 vestigated and from them the general oudine of the behaviour obtained, in the account 

 which follows attention will be almost confined to the three wild samples listed under 

 (2), (3) and (4), owing both to the obvious advantages of wild over horticultural 

 material but also to the spread of chromosome numbers displayed, which recalls in 

 some ways what has already been noted in Dryopteris Borreri. 



Chromosome counts of all these strains showed that whereas the ItaHan material, as 

 expected, was of a low chromosome number with 2n = c. 60 (actually 58, see Figs. 174, 

 175), which may be interpreted as diploid, the Uganda material was tetraploid with 

 2n ^c. 120 (cf Fig. 176^). It is of some interest to have traced a tetraploid to a known 

 geographical locality. On the other hand, 'var. albolineata' is triploid with yi = c. 90 

 (Fig. 176c), and so is 'var. albolineata cristata' . 



All forms of Pteris cretica give an exquisite quality of fixation readily, as may perhaps 

 already have been seen in Fig. 173 and others. It is also fortunate that sixteen-celled 



174 



