CampanulatcE. 301 
one, Vaccinioideae will readily find a place among Inferse. Signi¬ 
ficant in this connection are, first, the prevailingly racemose 
inflorescences, and the absence of any tendency to the “umbelli- 
floral” arrangement, (chap. VII). Second, the introrsely dehiscing 
anthers are approximated to each other and to the style, in definite 
relation to pollination. These two characters point to affinity with 
the Campanal Stock, as opposed to the Rubialian, in which the in¬ 
florescences are umbellifloral and the anthers separate, if not 
prevailingly divergent. Lastly, this suggestion is not discouraged 
if we compare the habit and general facies of the Thibaudieae 
section of Vaccinioideae with that of the shrubby and arborescent 
members of Campanulaceae. 
This line of reasoning, dependent, we repeat, upon the assump¬ 
tion that the epigyny of Vaccinioideae is an ancestral character, 
leads us in this way to regard the group as representing the 
Heteromerae of the campanal stock. In other words, Vaccinioideae 
will represent the transition, from the aspect of economy of the 
androecium, between the diplostemonous Archichlamydeae and their 
haplostemonous sympetalous progeny—a link which we have failed 
to find in the case of the hypogynous Sympetalae, the Bicarpellatae 
(chapter IV). In the case of Rubiales, a similar link with Archi¬ 
chlamydeae is not of similar weight, as isomery of the androecium is 
already accomplished in Umbelliflorae. 
Against all this there is, however, an objection which is regarded 
by most systematists as paramountly serious, namely, the close 
resemblance between Vaccinioideae and other groups of Ericaceae 
in respects other than that of the relative position of the ovary; 
and in modern systems the ericalian affinity is universally recognized. 
This side of the case has been emphatically stated recently 
by a well-known British systematist, 1 substantially in the 
following terms: “If the Vaccinioideae are to be removed from 
“ the Ericaceae, the Arbutoideae must go with them, on the ground 
“ of the specialized characters of the anthers, fruit, and seed . . . . 
“ the result would be a most unnatural separation, because the 
“ Arbutoideae and Vaccinioideae together have many characters in 
“common with the rest of the family. Hence,” he adds, “this 
“ suggestion would result in chaos! ” The “ suggestion ” was, never¬ 
theless, adopted by no less a philosopher than Lindley, who says : 
“ It is usual to station these plants (Vacciniaceae) with Heatlnvorts 
1 Dr. Moss, in the lectures referred to in Chapter VI, to whose 
views as a philosophical systematist the writer has from time 
to time been materially indebted. 
