230 Sargent.—Fragments of the Flower Biology 
The Origin of Honey-eating. 
Many times, early in the morning, in spring, when a certain Pear-tree is 
in blossom, I have seen small birds ( Zosterops gouldi) sipping dewdrops 
from the leaves of the said tree, and, just occasionally, dipping into a 
blossom for a sip of sweetness. While in Albany some years ago I saw 
the same bird taking dewdrops from the foliage of Adenanthos cuneata , 
disdaining apparently the sweeter moisture of the flowers. After a shower 
a large drop of water remains at the eye of Templetonia retusa flowers. 
These observations seem to me strongly suggestive of the origin of honey¬ 
eating in birds. How profoundly what was perhaps an accident—the 
mistaking of a nectar-drop for a dewdrop—in the far distant past has 
modified the present face of nature ! 
Variation. 
Three well-marked instances of variation have come under my notice. 
The labellum of Caladenia dilata , R. Br., is usually deeply fringed ; but 
occasionally a flower occurs fringeless. In one locality, Mt. Bakewell,I have 
never yet found a single specimen with a fringed labellum ; all are fringeless. 
I have seen no intermediates. The change is abrupt and striking. The 
inner rows of the involucral bracts of Helichrysum Lawrencella , F. Muell., 
are petaloid and usually bright pink. A solitary plant with pure white 
bracts is occasionally found. On the slopes of Mt. Bakewell white-bracted 
plants are common, and often form small communities. When such a com¬ 
munity is bordered by a pink-bracted one, intermediates of various degrees 
occur. In this locality pure whites seem the most vigorous. I believe both 
the above species are frequently self-pollinated, though with insect aid. 
Drosera macrantha , Endl., is usually white flowered, but in one locality 
I have found pink-flowered plants fairly common, intermixed with the white 
ones. The pinkness varies.in shade from the merest suggestion to pale rose. 
In the year 1913 thirty-two white-flowered plants produced 39 per cent, of 
fruits from their flowers, and 36 per cent, of the flowers of sixteen pink- 
flowered plants were fertile. Selecting several different sixteens from 
amongst the thirty-two, I found percentages varying between 33 and 48 ; 
so that the percentage for the pinks falls well within the limits of observa¬ 
tional error. My conclusion is, therefore, that a pink tint to the petals had 
no selection value that year in the locality in question. In the years 1912 
and 1914 I managed to examine a few plants of each variety, and on both 
occasions the white flowers showed a higher percentage of fertility ; so possibly 
a pink tint is detrimental to these flowers. This raises the question of the 
value of colour as an attractive agent. All the definite evidence I have col¬ 
lected is negative. Under suitable conditions Marianthus lineatus blossoms 
develop their sweet odour quite strongly; and then they are powerfully 
