74 SYSTEMATIC ARRANGEMENT OF THE SPECIES 
or less crenate or fimbriate scale. Towards the base the scales are always “adnate in the middle,” 
or, properly speaking, attached to both sides of the adnate filament. Their bases usually connect 
with one another, forming inverted arches. 
In the following species these scales are wanting: C. grandiflora and C. prismatica of South. 
America, C. hyalina of Asia, C. Californica, and ©. Sandwichiana. In C. Californica the inverted 
arch alone is present, entire or fringed ; in the others I find no trace of scales at all. 
These scales are evidently lateral dilatations of the lower (attached) part of the filaments, 
perhaps of the character of stipules, as Professor A. Braun suggests ; or they are a sort of stamineal 
crown, attached at base to the corolla, but not a duplication of the same. 
The presence, form, and size of the scales furnish some of the best characters in this genus, but 
they are not entirely reliable ; and while in some species they are very constant, in others they are 
found to vary considerably. It is doubtful whether a really scaleless form of C. Huropwa exists ; 
C. Californica, usually without scales, seems to occur also in a variety with scales. 
The ovary and pistils are more reliable for the determination of species, just as they furnish the 
most important characters for the distinction of the sections. The walls of the ovary are of equal 
thickness throughout, or they are thickened towards the base of the style (furnished with a 
stylopodium, as I formerly designated this form). The ovary is smaller than the tube of the corolla, 
or it fills its whole cavity, or even protrudes from it. The styles are subulate or terete, thick or 
capillary, and very constant in these differences. Their length, however, is variable; and this 
character, so much relied on by Choisy in the subdivisions of this genus in his Monography and in 
romus, is of secondary importance, as the same species sometimes occurs 
with short or with long styles, and as the styles, included at first, often become exsert [458 (8)] 
with age. The direction of the styles in the flower and on the fruit furnishes a tolerably 
good character. 
The position of the dead corolla on the capsule has already been mentioned as a pretty reliable 
specific character. The shape and even the texture of the capsule also ought to be noted, though in 
several species (C. Huropwa, for example) its form is quite variable. 
e number of seeds which ripen in each capsule furnishes no distinction, though the species 
with very crowded flowers, and some others with loose flowers also, often develop only one or few 
seeds. The shape and surface of the seed ought to be studied more, and will yet, it is believed, help 
to distinguish some species. 
As almost all the characters enumerated above are subject to more or less variation, it is neces- 
sary to base the diagnosis of a species on a combination of a number of characters; but as the value 
of these characters is necessarily differently estimated by different botanists, some will consider as 
well-marked species what others will look upon as mere varieties. 
The different species often seem to have a predilection for certain plants, or families of plants, 
for their sustenance; and I have myself at times thought I discovered an influence of the mother 
plant (or, better, nursing plant, nurse) on the form and development of the parasite. But I have 
become fully convinced that this influence is very limited, and probably goes not even farther than 
the influence of different kinds of soil and manure would go with any other plant. If some species 
seem very constantly to prefer certain plants to others (C. Europea, Urtica dioica ; C. Epithymum, 
Calluna vulgaris or Genista sagittalis; C. chlorocarpa, Polygonum; C. Gronovti, Cephalanthus ; 
C. lupuliformis, Salic ; and, the most marked example, C. Epilinum, the flax fields), it is probably 
because the kind of soil, the humidity or dryness, the shade or sun, and all the circumstances which 
suit the nurse, also agree best with the parasite. On the whole, succulent herbaceous dicotyledo- 
nous plants suit them best as nurses; some few species prefer low shrubs or semi-shrubs; and most 
of the Monogynelle and a few others affect larger shrubs and trees, of course penetrating only the 
tender bark of the smaller limbs. 
