INTRODUCTION. XXV- 
may be known from arvensis by that distinction among’ 
others. Their colour deserves attention if not owing 
to pubescence. The glaucous hue of rubrifolia contri- 
butes to distinguish it from canina, and laxa from lu- 
cida. The shape of the leaflets, unless remarkable, can’ 
rarely be employed; ovate varies to oval; and orbicu-— 
lar to ovate; lanceolate and retuse are the most con- 
stant forms under which the leafletsappear. The latter ' 
shape would alone be suflicient to separate sericea from 
the species near which it stands. Concavity is unim- 
portant in most instances. Yet it is a remarkable fea- 
‘ture in &, lutea, whose leaflets are hollowed like the 
bowl of a spoon, and confirms the affinity of that spe- 
cies to rubiginosa, in which the same peculiarity exists. 
The rugosity of the leaflets will frequently furnish good. 
specific characters. It distinguishes Villose from Ca. 
nine, acicularis from the rest of its division, and cin- 
namomea from majalis. The shining leaflets of feror 
are very unlike the opaque ones of kamchatica. Single 
and double serratures can only be made use of under 
particular circumstances. Generally they are as vari- 
able as any thing; there are states of canina which 
would puzzle the most practised eye to decide whether 
the leaflets were simply or doubly serrated. Myriacantha 
is asserted by Mr. Woods to have double serratures: in 
all my specimens they are simple. A plant of that va- 
riety of rubiginosa which Mr. Woods has called Bor- 
reri, one year produced all its leaves with simple ser- 
ratures; and the next with compound ones. Yet I be- 
lieve the double serratures of involuta will prevent its 
being mistaken for spinosissima. 
_ The inflorescence varies in different species from the 
most, simple to the most compound form. It however 
d 
