98 
the total up to fifteen. Bentham and Hooker (l.c.) reduced 
this number to seven, in which they were followed by Prantl 
and Kundig in Engler & Prantl, l.c. In regard to the classifi- 
cation of the Papaveraceae by the last-mentioned authors, it 
should be noted that their arrangement differs widely from any 
other proposed in that they interposed the ‘“ Papaveroideae ”’ 
between the “‘ Hypecoideae’’ and ‘‘ Fumarioideae.” As thus 
classified we have at the beginning of the family two genera 
with dimorphic petals and four stamens, in the middle a large 
group of genera (the Papaveraceae proper) with uniform petals 
and numerous stamens, and at the end the ‘“‘ Fumarioideae ”’ 
with very zygomorphic flowers, dimorphic petals and six stamens 
united into two bundles. This classification seems unnatural, 
for whatever status the groups take, the “ Hypecoideae”’ and 
** Fumarioideae ’’ should undoubtedly form one principal group 
with further subdivision, and the ‘“ Papaveroideae ”’ a second 
principal group to be dealt with likewise. This course is adopted 
here, after a careful examination of all the species of the two 
families represented at Kew, and the present account deals with 
Fumariaceae, divided primarily into two subordinate groups, 
_ Hypecoideae and Fumarioideae. 
In regard to the number of genera which may be recognised, 
it will be seen that there is considerable divergence of opinion 
in general systematic works. But authors of local floras and 
smaller systematic papers have as a rule taken the genera in 
a more limited sense. Mr. Pugsley* has recently published a very 
careful revision of the genus Fumaria, from which he has separated 
and revised the genus Rupicapnos, at the same time pointing 
out the desirability of recognising the genera Platycapnos, 
Bernhardi, and Discocapnos, Cham. & Schlecht., all of which had 
been reduced to Fumaria. I agree with this recommendation, and 
in order to apply the same principles to the other genera of the 
family, the following classification is proposed. In the case of 
the smaller genera, I have also taken the opportunity of giving 
the records showing their distribution, which is presented in 
more graphic form on the accompanying maps. No new genera 
are proposed in the present paper. 
Geographical Distribution Fumariaceae are most abundant 
in the north temperate parts of the Old World, especially around 
Mountains. Central and South America, Tropical Africa’ (except 
the high equatorial eastern mountains), South India, Malaya and 
Australasia are entirely destitute of Fumariaceae excepting weeds 
of cultivation such as Fumaria officinalis. South Africa, how- 
ever, has four small endemic genera, widely separated geogra- 
phically from any other members of the family (see map, p. 105). 
* W. Pugsley, ‘“‘ A Revision of the genera Fumaria and Rupicapnos’”’ 
in Se Linn. Soc. Bot. xliv. 233-255 (1919). 
