24. ROSACER. 507 
or changcability of species, even those of ‘“ primary value,” that the 
two reduced species are now not placed as varieties of their con- 
geners in the same group, but each of them is changed into a 
variety of another species in a different group. Borreri has become 
a variety of canina, the sole species of group 4; while Jundzilliana 
has sunk into a variety of tomentosa in group 2. Cryptopoda is 
retained in group 3, but there sinks into a variety of pulverulenta, 
an aggregate species first appearing in 1869. 
Other evidences are to be found in the ‘ Monograph,’ shewing 
that the combination of varieties into species of “ primary value” 
is really much a matter of whim or fancy. Pretty full diagnoses 
are given for the species, which look very important and imposing 
in their latin words ;—and then succeed descriptions of varieties 
which decidedly contradict the diagnosis, through the variety of 
one species assuming some of the distinctive characters attributed 
to the other species. Contrast the diagnoses of rubiginosa and 
micrantha, for instance, and then compare the descriptions of their 
varieties with the words of their specific diagnoses. It will be seen 
that the diagnostic characters of the one species appear in the 
varieties of the other species. 
Such uncertainties and inconsistencies are not here imputed as 
faults in Mr. Baker’s workmanship. Their true significance is 
quite different. They really shew that the various forms of our 
wild roses are so connected by the interchange and crossing of 
technical characters, only imagined to be diagnostic between them, 
that the book-species really depend on an arbitrary preference given 
to this or to that set of characters, as indicating affinity, and as 
necessitating union or severance. 
With the omission of Jundzilliana and cryptopoda, two spe- 
cies which I was unable to understand at all, and the reluctant 
acceptance of Borreri, the species of the Synopsis on preceding 
pages of this volume are essentially the same with those of 
Mr. Baker’s Review. Substituting pulverulenta for Borreri, they 
would be made to correspond with those of the new Monograph, 
although not absolutely so in respect to all included varieties. 
Probably much less so in respect to the recorded localities, on 
faith of which the formula for their distribution was filled in. 
Anticipating now that the names and arrangements of the 
Monograph will be generally accepted by English botanists, I wish 
to add a tabular arrangement of the provincial distribution, as 
nearly similar to the one taken from ‘The British Rubi’ as the 
different treatment by the two Authors will allow it done. I shall 
seldom venture to add any province besides those in which 
Mr. Baker has cited localities, either in the Monograph or else- 
where and recently. The provincial nos. will thus be few, and 
blanks be numerous; for Mr. Baker has accomplished in the 
topographical department of his British Roses nothing like the 
