624 DR. M. T. MASTERS ON THE PASSIFLORACEZÆ. 
inseparably connected with the work done by those particular plants; hence their value 
is not only physiological, but also analytical, because they furnish good marks whereby 
io discriminate one species from another. In a purely morphological sense thése cha- 
racters may often appear trivial; but if they are associated with a difference in physio- 
logical function, it cannot be said that they are unimportant. Itis sufficient to mention, 
by way of illustration of this class of characters, all those diversified arrangements for 
favouring or preventing cross fertilization. Thus, in the case of the Passiflore, it has been 
shown that the arrangement of the series of threads and membranous rings constituting 
the corona is of great value in separating species, or subsections of the genus, one from 
another. These organs are “acquired,” they are formed late, and they have a special 
office in favouring cross fertilization, and in impeding or preventing self-impregnation. 
I call characters derived from such organs as these “ special physiological characters," not 
only because they are peculiar to certain plants, but also because their office is special. 
They are not essential organs (fertilization can and does take place without their aid); 
but they modify the action of the essential organs, and make them perform their office 
in a peculiar manner. The remarks that have been made on acquired characters apply 
then to those which are here termed special physiological. 
It is not requisite in this case to say any thing concerning characters dependent on 
teratological phenomena or on those founded on varying geographical distribution of 
species. It would occupy too much space to treat of these subjects at length in such a 
communication as this; but, by way of summing up, I may add that, in estimating the 
value to be attached to certain characters or points of distinction, it is necessary to con- 
sider the purpose for which they are required. If the object be syuthetical—if we are 
seeking points of resemblance, so as to group together a large number of forms into onc~ — 
or more large aggregates—we must, in the first instance, lay stress on the congenital 
characters, as serving to bind together the greatest numbers, and as being the most 
constant, the least variable, and the most physiologically important; then we must conf 
sider those dependent on frequency of occurrence and special physiological office, aft 
wards such others as may be available or forthcoming. If the purport of our inv E 
gation be analytical; if discrimination be our object, the special physiological characi 4) 
first demand attention, then those which have the merit of frequency and invariabilit, 
and lastly those that are congenital. 
It stands to reason that if a character be at once numerically important, congenital, 
and of high physiological value, it will be most valuable for synthetical purposes: for! 
instance, in the large group Parietales, including Exogens with parietal placentation, the 1 
. presence of albumen is a very constant feature (except in Cueurbits); hence then, asa | 
character, albumen, in such a case, is numerically important, it is congenital, and physi- - 
ologically of high importance. It is hardly necessary to add that the systematist cannot 
always aet up to his own standard. Even the most experienced and sagacious botanists i 
are not unfrequently inconsistent, often must be so from defective evidence. Individual 
Cases have to be dealt with on their own merits, and consistency often has to give place 
to convenience or expediency. | | 
From what has been previously said, it would seem as if the proper position of Passi- 
| 
