322 



T. G. TUTIN 



However much time one may waste in trying to define a genus or to standardise 

 generic criteria, the personal opinion of individual taxonomists is bound to come in. It 

 is in fact remarkable that so large a measure of agreement about generic limits exists. I, 

 personally, like fairly "small" genera for two reasons : first there is a greater chance 

 of them representing natural, that is to say phylogenetic units, and, second if they are 

 reasonably homogeneous they are more easily recognisable and therefore more con- 

 venient. I have not talked about convenience before but, in the absence of a reliable 

 phylogenetic system, convenience is perhaps the most important thing to aim at in a 

 general purposes classification. 



A consideration of some of the British representatives of the tribe Scirpeae of the 

 Cyperaceae may help to illustrate the use of a number of these criteria and also demonstrate 

 the importance of the personal factor. 



Hooker (1884), and more recently Hutchinson (1934), group the British species of 

 this tribe in three genera, Eleocharis, characterised by the persistent style-base, Eriophorum, 

 with generally numerous bristles which elongate and become silky in fruit, and Scirpus, 

 with neither of these characters. The flowers are closely similar in all the species, apart 

 from the bristles, which vary from none to many and from long to short, and the occurrence 

 of the swollen style-base in some species. On the other hand the inflorescence and more 

 especially the vegetative parts show a considerable range of structure. 



If one adopts the broad Hooker view of Scirpus, covering plants as different vegeta- 

 tively as S. setaceus, S. lacustris, and S. sylvaticus and others, such as S. pauciflorus, 

 which only differ from Eleocharis in the absence of the swollen style-base, it is logically 

 impossible to separate Eleocharis and very difficult to maintain the distinctness of Erio- 

 phorum. Thus Lid in Norsk Flora (1952) recognises only Eriophorum and Scirpus. If on 

 the other hand the nature of the inflorescence and vegetative and anatomical characters 

 are used, Scirpus sensu Lid can be divided into eight genera each of which can be recognised 

 at sight. 



The c\tological evidence is less complete than could be desired and some counts are 

 open to doubt, largely owing to the technical difficulties of the group, but using what 

 appear to be reliable figures we find the basic number 29 for Eriophorum and for Scirpus 

 (sensu Lid) 10, 11, 13, 21, 29, 31 and 43. The following table shows the haploid chromo- 

 som.e numbers of seven of the eight genera into which Scirpus is divided, using vegeta- 

 tive and inflorescence characters, in the Flora of the British Isles (1952) : 



Table 1 



Genus 



Basic chromosome number 



Trichophorum 



13, 29 



Eleocharis 



8, 10, 19, 23 (perhaps aneuploid) 



Scirpus 



31, 43, 55 (perhaps aneuploid) 



Holoschoenus 





Blysmus 



11 



Schoenoplectus 



21 



Isolepis 



13 



Eleogiton 



10 



It will be seen that the c>1:olog>', as far as it goes, supports the genera based on 

 non-floral characters on the whole, though the present classification is not altogether 

 satisfactory. The haploid number 29 in Trichophorum refers to T. alpinum which has 

 been put in Eriophorum, to which genus it would seem properly to belong. The size of 



