VOL. XVIII. (3) NOTES ON HELOSCIADIUM 235 



shade, and the permeabiHty of the soil ; perhaps all are, 

 though I think not. But what is abundantly clear is the 

 desirabiUty of growing on a large number of different forms 

 in one garden, in order to test the question of their funda- 

 mental identity, especially if the forms are gathered from 

 widely different localities. 



My behef is that almost all the named varieties of H. 

 nodiflorum are simply states due to difference of soil, water 

 supply, and cover. They should, therefore, be altogether 

 excluded from our British hsts, or, if mentioned, should be 

 lightly and summarily passed over as mere forms due to 

 surrounding circumstances. 



But there is one exception, viz., var. longipedunculatuni 

 F. Schultz, which may even possibly deserve to be called a 

 species. Schultz, indeed, has raised it to specific status as 

 H. palatinwn, though I am not sure that his H. palatinum 

 is identical with his var. longipedunculatuni ; but, in my 

 opinion, such a procedure depresses the rank of species in- 

 conveniently. The variety does not differ from type in 

 characters which really have specific value. 



Var. longipedunculatuni F. Schultz, then, is a real variety, 

 and not a " form " oi H. nodiflorum. Its best known locali- 

 ties are GuUane or Guillon Links, Haddingtonshire, and 

 Duddingston Loch, Midlothian, but it is by no means hmited 

 to these places. Herbaria contain many specimens of it, 

 more or less disguised, from widely scattered localities, in 

 Suffolk and Cambridge, but I beheve that, in spite of their 

 disguise, they are all so many forms of one thing, probably 

 produced simply by surrounding circumstances. A very im- 

 portant feature of this variety is that in certain circumstances 

 it closely simulates H. repetis, far more closely, in fact, than 

 any form of type H. nodiflorum does. English botanists, in 

 fact, have talked of H. repens for many years : they have 

 apphed the name to small, compact, strongly-rooting forms 

 of type H. nodiflorum which are simply caused by the shrink- 

 age of water in a normally wet locahty ; they have less fre- 

 quently given the name, though with much more reason, 

 to a form of var. longipedunculatum. One of my chief objects 

 in this paper is to distinguish the latter plant from true H. 



