THE POLYPLOID SERIES IN OSMUNDA 

 was a diploid plus i, i.e. with 45 chromosomes instead of 44 (see Fig. 27c), and two 

 were diploid plus 2, i.e. with 46 chromosomes. From the root-tip counts there is no 



reason to think that any of the remaining diploids which had not yet become fertile 



would show a greater degree of cytological unbalance 



than this, and none was deficient in chromosomes. 



It therefore seems certain that of all the prothalli 



formed, only those deviating in not more than two 



chromosomes from the haploid or diploid state were 



able to complete their life history. In other words 



natural selection, acting on the single character of 



fitness to reproduce, had at one blow eliminated all 



but a tiny minority of the prothalh brought under its 



influence. 



Natural selection is evidently acting in this case in 



a capacity which is the exact opposite of a progressive 



evolutionary influence; it is clearly a powerful force 



tending to eliminate aberrations and to maintain the 



stability of the species unchanged. Had the autopoly- 



ploid series in Osmunda originated in nature, as it could 



possibly have done, it therefore seems probable that it 



would not appreciably have aflfected the evolution of 



the species, for the tetraploid seems unhkely to be 



able to compete with the diploid in the wild state, 

 and the triploid would be eliminated quite rapidly by 

 its sterility. 



Disappointing as this conclusion may perhaps 

 appear, the facts described have a very considerable 

 comparative value which can at once be utilized. 

 Looking downwards to the Bryophyta it is clear that 

 the Osmunda series provides a very close parallel to 

 the situation induced in many of the mosses by the 

 Marchals, F. von Wettstein and others. In both groups polyploidy can be obtained at 

 will as an inevitable consequence of induced apospory, and in both the series is cut 

 short by sterility when a tetraploid gametophyte has been reached. Only in the 

 exceptional case of species hybrids, notably that between Funaria and Physcomitrium, 

 has an octoploid gametophyte been achieved, but to this there is as yet no parallel in 

 the ferns. 



Looking upwards to the Flowering Plants it is clear that the cytological manifesta- 

 tions of polyploidy in Osmunda are of exactly the same type as in those Dicotyledons 

 and Monocotyledons in which they were first elucidated; multivalents are as well 

 displayed as in Datura, pachytene pairing closely resembles Lilium or Tulipa. There is 

 therefore direct evidence to justify the extension to the Pteridophyta of a type of inquiry 

 first founded on experience in the higher plants. This in itself is an encouraging 

 beginning. 



Fig. 28. Leaves from two of the 

 progeny from triploid Osmunda 

 showing leaf aberrations. Both 

 plants were three years old at the 

 time the leaves were taken, but the 

 left-hand plant remained perma- 

 nently in the juvenile stage. For 

 further description see text. 

 Natural size. 



42 



