1920] Fernald,—Blackberries of New England 187 
but one of its supposed parents reaches its eastern limit in eastern 
Massachusetts (or possibly southern Maine), at least 200 miles across 
the Gulf of Maine from the nearest point of Nova Scotia. Again, 
R. tardatus (p. 83), treated as a hybrid of R. flagellaris (R. procumbens) 
and R. setosus and cited as if found only at Kennebunk, Maine, is a 
dominant shrub of boggy thickets and lake-margins on Prince Ed- 
ward Island and Nova Scotia and it is characteristic of some bogs 
and peaty shores of central Cape Cod. Yet of its alleged parents, 
R. flagellaris (even in its most inclusive sense) is not known east of the 
Kennebec valley and R. setosus is quite unknown on much-explored 
Cape Cod. Is it not, then, somewhat strange, if these are no more 
than local and very recent hybrids, that they should abound over such 
wide areas and hundreds of miles away from one or both of their 
supposed parents? 
It is, in fact, very difficult to make out the principle by which the 
hybrids of Brainerd & Peitersen's treatment are differentiated from 
the true species. R. elegantulus (p. 37) with “Pollen about 70% 
imperfect" and a restricted range (the uplands of New Hampshire 
and Vermont), and R. vermontanus (p. 39) with “ Pollen about 85% 
imperfect" and a distribution said to be confined to New Hampshire 
and Vermont, are treated as true species. But R. frondisentis (p. 
63) of similar range and with “Pollen about 10% imperfect," a 
plant with seedlings which “are very uniform and seem to breed 
true to the type,” and R. abbrevians (p. 65) again of similar range, 
and R. permiatus (p. 69), extending from New Hampshire to New 
York and New Jersey, the former with “Pollen about 10% imper- 
fect” and seedlings which “ vary very little from the mother plant,” 
the latter with “ Pollen about 50% imperfect" and seedlings which 
“do not revert to the parent types," are treated merely as hybrids. 
If these characteristic and easily recognized plants are indeed hybrids 
they are notable refutations of the much overworked theory, that 
hybrids have imperfect pollen and do not breed true. 
A further refutation is found iu the fact, that two of the univers- 
ally recognized species, admitted without hesitation by the present 
authors and by every other competent systematist, have as poor 
pollen as is found in the genus. These are the “Thornless Black- 
berry," R. canadensis (p. 35), ranging from Newfoundland to Wis- 
consin and the mountains of North Carolina and Tennessee and the 
"Running Swamp Blackberry" (which often grows on dry sand 
plains), R. hispidus (p. 43), with an almost equally broad range, 
from Nova Scotia to southern Ontario, Michigan and North Carolina, 
the former with “Pollen about 85% imperfect," the latter with 
“Pollen about 90% imperfect." Furthermore, R. frondosus (p. 
31), not treated as a hybrid, has seedlings which “show quite a range 
of variation as to shape of leaves, serration of leaflets, etc." 
In the discussion on p. 11 the statement is made, that “Seeds from 
the selfed flowers of a number of suspected hybrids have been grown 
