252 MR. H. W. PUGSLEY : A REVISION OF THE 
tightly enclosed within the connate petals. As in this rudimentary 
condition the characteristic features of the flower are more or less lost, it is 
desirable in the consideration of species to have regard to perfect flowers, 
with the corolla fully coloured and winged, and provided with a nectary. 
In these perfect flowers the lower petal is generally deflexed and free, so that 
the two outer petals are apically distant, in nearly every species. Hammar’s | 
definition that this is so only among the Agrariæ, while in the Capreolate and 
the Officinales the lower petal cuheres to the others until separated at the base 
by the swelling fruit, cannot be accepted, for apically distant outer petals are 
frequently to be seen in good flowers in all three classes, as may be inferred 
from the observations of Jordan and other authors; and in all of them the 
lower petal eventually becomes detached at the base as the fruit develops. 
In some of the Agraria, however, the lower petal seems more caducous than 
in most other species. 
A feature of the upper petal, not noticed by Hammar, but apparently 
constant throughout the genus, is that while the apical wings are formed by 
the margins of the petal at the very apex and for the greater part of their 
length, towards the base of the petal they leave the margin and are finally 
produced in a blunt lateral ridge. The tube of the upper petal may be 
regarded as laterally or dorsally compressed in proportion as this lateral ridge 
is distant or close to the edge of the petal. 
When good flowers are present in exsiccata, a difficulty sometimes arises 
owing to the form of the corolla being obliterated in the process of drying, so 
that its dorsal or lateral compression, the shape of its spur, and the direction 
of the margins or wings surrounding the apical keels of the outer petals are 
not readily seen. In the living plant these differences are often very marked, 
the acute, laterally compressed corolla, with large, curved spur and narrow, 
strongly reflexed wings to the upper petal, seen in such a species as 
F. capreolata, being widely divergent from the obtuse, dorsally compressed 
corolla, with smaller, straighter spur and broad, erect-spreading wings, of 
F. agraria. In adequate exsiccata the direction of the wings and the form 
of the spur can usually be fairly judged after some experience, but pressed 
flowers are often misleading respecting the compression of the corolla-tube. 
A feature of the corolla that does not appear to have been sufficiently 
appreciated is the constancy of its colour and of the dark marking about its 
apex in almost every species throughout the genus. So far as is known, 
there is no Fumaria (with the possible exception of F. parviflora, one variety 
of which (indicoides) seems to possess essentially pink corollas) in which the 
colour of fully developed flowers may be either pink or white, although pink 
flowers may become whitish when depauperate or shade-grown, and white 
ones eventually entirely suffused with pink or red after fertilization. But 
the essential colour, which is best seen in the later bud stage or as the flower 
opens, seems to be invariably constant. 
