PART IV. | GENERAL RETROSPECT. 587 
those which are least visited are those in which self-fertilisation 
is best insured.! 
The comparative merits of cross-fertilisation and self-fertilisation 
can only be rightly considered when the actual insect-visits that the 
plant receives are taken into account ; and only the utter neglect 
of this precaution can explain how Hildebrand and Axell came 
to directly opposite estimates of the value of the two modes of 
fertilisation. For Hildebrand in his Geschlechtervertheilung ar- 
ranges the mechanisms of flowers according as they hinder self- 
fertilisation, and seems to consider those the most perfect in which 
self-fertilisation is least possible; while Axell tries to show that 
the highest and most perfect Phanerogams are those which 
regularly fertilise themselves.” 
Both views are only partially correct ; the truth lies between 
them. The law, proved by general considerations in the Introduc- 
tion and by many detailed facts in the third section of this book, 
that cross-fertilisation is better for a plant than self-fertilisation, 
but that self-fertilisation is infinitely better than no fertilisation at 
all and consequent sterility, must modify essentially Hildebrand’s 
conception of a law of the avoidance of self-fertilisation. And it is 
evident at once that Axell’s idea of a single path towards perfection 
in the evolution of Phanerogamic flowers is altogether untenable 
when we review the insurance of cross- and self-fertilisation in 
connection with the insurance of insect-visits. 
The oldest Phanerogamic flowers which adapted themselves 
for the transport of their pollen by insect-agency certainly pos- 
sessed those characters by which insects are mainly attracted— 
namely, conspicuousness, perfume, and honey—to so small an 
extent that as a rule their insect-visitors were not numerous 
enough to insure cross-fertilisation. Under these conditions it 
must have been advantageous for the organs of both sexes to be 
united within one flower so as to admit the possibility of self- 
fertilisation. In accordance with this is the fact that while the 
Gymnosperms have diclinic anemophilous flowers, the great 
majority of entomophilous Phanerogams are monoclinic. 
But as soon as in any entomophilous plant increased conspicu- 
ousness, or perfume, or food-supply had so far multiplied insect- 
visits that cross-fertilisation took place regularly and the possibility 
of self-fertilisation became quite useless, then the possibility of 
1 Compare Rhinanthus, Lysimachia, Euphrasia officinalis, Rosa, Rubus, Epilobium, 
Geranium, Malva, Polygonum, Stellaria, Cerastium, Veronica, Hieraciwm, Senecio, ete. 
2 Cf. ‘Historical Introduction,” pp. 12, 20. 
