INTRODUCTION TO THE PROBLEM 



it must have occurred more than once, since different species are polyploid in diflferent 

 degree, while in others it may have so far escaped detection owing to an insufficient 

 number of species available for study. That a proportion, possibly a high proportion, 

 of these cases are allopolyploids of the type of Nasturtium uniseriatum seems highly 

 probable both from experience within the family (e.g. Erophila, Raphanobrassica, etc.) 

 and from the sum of evidence from other flowering plants. 



There are, however, undoubtedly other methods of species formation in operation. 

 Thus, polyploidy apart, the diploid species oi Biscutella of the Laevigatae section appear at 

 present to be evolving by purely genetical means. Many of them have a distinctive 

 morphology, which is retained in cultivation, and a characteristic ecological habitat : 

 the high alpine B. mollis, endemic to Austria, and the French B. Lamottii confined to 

 extinct volcanic ash heaps of Tertiary age in the Auvergne, being good examples. 

 Some of the morphological and probably also physiological characteristics of these 

 species (or ecotypes as they may perhaps more properly be called) are undoubtedly 

 adaptive and enable their possessors to colonize successfully types of habitat which are 

 completely closed to the parental forms which, in both cases, are almost certainly still 

 present in the same geographical area though far removed from the actual sites. This 

 at once confronts us with something closely resembhng simple Darwinism, but the 

 species in question are all still sufficiently akin to be interfertile if brought artificially 

 into contact and have completely regular chromosome pairing in their hybrids. 



Experience has, however, shown that speciation by genetical means uncomplicated 

 by other cytological changes is either relatively unusual or only a passing phase. 

 Sooner or later in most species, and probably also at some future time in Biscutella, 

 internal changes in the chromosomes set in so that homology is lost. Chromosome 

 pairing in hybrids then ceases to occur, and we have the type of sterility barrier which 

 is the taxonomist's greatest ally in determining natural specific boundaries. We know 

 very little, unfortunately, about the mechanisms by which sterility barriers due to a 

 loss of chromosome homology are brought about.* The pioneer in drawing attention 

 to the need for such knowledge was again Bateson in 191 3. A beginning has, however, 

 been made by cytogenetic work carried out largely in America, Russia, and to a lesser 

 extent in Great Britain, on a wide range of experimental plants and animals in which 

 the effects on chromosome pairing of visible lesions such as fragmentation, translocation, 

 segmental interchange, etc., have been studied. The exact role of these processes in 

 species formation is less easily demonstrated than are the other methods previously 

 discussed, though the work on the species of Drosophila and Crepis mentioned on p. 5 is 

 sufficient indication that in certain cases their importance is considerable. In the Cruci- 

 ferae there is practically no work directly devoted to such studies, but the mere observa- 

 tion that species hybrids here, as elsewhere, are more commonly sterile than fertile is 

 sufficient ground for believing that here also an important effect of intrachromosomal 

 structural changes must be envisaged. 



A different type of nuclear upheaval to which the intrachromosomal changes just 

 mentioned may perhaps provide a clue, is that of aneuploidy (sometimes termed 



* An outstanding exception is to be found in the work of Tobgy and Gerassimowa on Crepis, sum- 

 marized by Babcock ( 1 947 a) . 



14 



