the distilled water were full of Bacteria, the others remaining 
barren. This was done to show that the liquid, although de- 
prived of its power of germination, is as capable as before of 
supporting the life of Bacteria. 
The results of the preceding experiments may be summed up 
as follows :—In sixteen experiments the liquids were subjected 
to the temperature of boiling at the normal pressure ; of these, 
eight were heated for 15 minutes, and all bred Bacteria ; six were 
heated for 30 minutes, of which two bred Bacteria ; two for an 
hour, both of which were barren. 
Of ten subjected to the temperature of ebullition at pressures 
not exceeding one inch, eight were barren. Both the liquids 
which were found to be pregnant lad been heated for 30 
minutes, one under a pressure of one-tenth of an inch, the 
other of one inch. 
In the twelve experiments in which the liquids were heated 
under pressures exceeding one inch, all were barren, although 
half oe them were subjected to that temperature for only 15 
minutes. 
It is unnecessary for me to draw any inferences from the 
preceding experiments ; it may not, however, be superfluous to 
point out that, although all the flasks heated above ror’ C. re- 
mained sterile, this fact affords no ground for concluding that any 
definite relation exists between that precise temperature and the 
destruction of the germinating power of the iiquid in question. 
All that has been shown is that the chance that such a liquid 
will breed Bacteria is diminished either by slightly increasing 
the temperature to which it is heated, or increasing the duration 
of the heating. Thus it appears to me quite probable that if a 
sufficiently large number of flasks were heated even to 102° C., 
some ofthem would still be found to be pregnant. 
University Coll., London, June 7 J. BURDON SANDERSON 
Fertilisation of the Pansy.—Ground Ivy 
THERE is one further point in the structure of Lola tricolor 
which is not mentioned by Mr. Bennett or by Mr, Hart, but 
which seems to confirm the theory of the former gentleman that 
V. tricolor, as distinguished from most other Violas, is fertilised 
by a small insect such as Thrips instead of by the proboscis of 
larger insects. 
Before saw Mr. Bennett’s paper, my attention had been called 
by Miss Dowson to the fact that whereas in the Sweet and Dog 
Violets, the circle of anthers presses close to the style all round, 
there is in V. ¢ricolor an opening between the two appendaged 
stamens. The use of this opening will evidently be to allow 
the small creature to enter in and craw! down the stamen to the 
nectary at the end of the appendage. This structure may be also 
seen in . cornuta, which seems to be fertilised in the same 
way. In V. “ricolor the opening is exactly opposite to the 
black streak, or guide-post, as Mr. Bennett has termed it. 
In V. cornuéa, although this black mark is notso evident, there 
is a distinct triangular mark pointing downwards exposed by the 
opening of the stamens. On each side of the style are two sets 
of hairs, looking like ‘* whiskers” to the scull-like crest of the 
style, on which lots of pollen rest. The small insect on enter- 
ing the flower can hardly help crawling into the cavity at the top 
of the stigma, for the entrance to the flower is almost completely 
blocked up by it. On emerging from it it would crawl over the 
top, which Mr. Hart mentions as seen in ¢ricolor, and which I 
also find in cornuéa, be guided through the hole by the trian- 
gular mark, and so find his way to the nectary. On emerging, 
covered with pollen, and entering its next flower, it will again be 
deluded into the cu/-de-sac in which the stigmatic surface is, where 
it will deposit its pollen, The details of the structure of the 
appendaged stamens, as contrasted with those of other Violas, 
fully bear out this view. ‘ 
As regards the English translation for the German Jdestduden, 
I would suggest to Mr, Hart that ‘‘pollenate” is an impossible 
word ; follen, pollinis, must give the verb to ‘* pollinate,” as 
fulmen, fulminis gives fulminate. But there isa greatadvantage 
in a word which speaks for itself, and, if the word ‘ be-pollen” 
offends scientific ears (Mr. Hart does not tell us why), would 
the literal translation of the German. ‘‘ to be-dust”’ be offensive ? 
If not, I think it would tellits own tale. The word ‘‘empollen” 
seems justified by evda/m, but the prefix generally means to 
place in or convert into, as in exthval, emprison, embed. Wence 
it would at least be ambiguous. 
The form of Ground Ivy mentioned by your correspondent 
S.S.D. grows here abundantly in several spots, seeds freely, and is 
remarkable for haying a much shorter style in proportion to the 
NATURE 
143 
tube of the corolla than the common form in which the style 
and stigmas protrude from the tube. F. E. KircHENER 
Rugby, June 15 
Mr. Kitchener having been kind enough to send me the 
above letter, I may, perhaps, be allowed to add a few additional 
notes. Since writing the former paper I have had the oppor- 
tunity of examining three other species of Viola, V. calcarata, 
elatior, and factea, all of which present a remarkable contrast to 
V. tricolor in a very curious point of structure. In V. tricolor 
the stigma is brought into close contact with the lowest petal by 
a very peculiar ‘‘knee” in the style, the effect of which is so 
completely to close up the central cavity of the flower as to 
render it extremely difficult for any large insect to insert its pro- 
boscis into the spur. In all the three species above-named, 
which I believe to be fertilised by bees, the style is nearly 
straight, so as to leave a considerable gap between the stigma 
and lower petal, quite large enough for the insertion of the pro- 
boscis of a bee. In none of these is there the least indication of 
the black triangular streak on the style which I take to serve, in 
V. tricolor, the purpose of guiding the Thrips to the nectary. 
The ring of anthers is also perfectly closed, as described by Miss 
Dowson in the case of the Dog and Sweet Violet, there being no 
opening for the admission of the small insect, as in the pansy. 
A striking difference in the form of the stigma also favours the 
same conclusion as to the mode of fertilisation. 
ALFRED W. BENNETT 
ON THE ORIGIN AND METAMORPHOSES OF 
INSECTS * 
VI. 
PoE metamorphoses of insects ‘have always seemed 
to me one of the greatest difficulties of the Dar- 
winian theory. In most cases, the development of the 
individual reproduces to a certain extent that of the race; 
but the motionless, imbecile pupa cannot represent a 
mature form. No one, so far as I know, has yet at- 
tempted to explain, in accordance with Mr. Darwin’s 
views, a life history, such as that of a butterfly, in which 
the mouth is first mandibulate and then suctorial. A clue 
to the difficulty may, I think, be found in the distinction 
between developmental and adaptive changes ; to which 
I have called attention in a previous article. The 
larvee of insects are by no means mere stages in the 
development of the perfect animal. On the contrary, they 
are subject to the influence of natural selection, and un- 
dergo changes which have reference entirély to their own 
requirements and condition. It is evident, then, that while 
the embryonic development of an animal in the egg may 
be an epitome of its specific history, this is by no means 
the case with species in which the immature forms have a 
separate and independent existence. If an animal when 
young pursues one mode of life, and lives on one kind 
of food, which subsequently, either from its own growth 
in size and strength, or from any change of season, alters 
its habits or food, however slightly, itimmediately becomes 
subject to the action of new forces: natural selection 
affects it in two different and, it may be, very distinct 
manners, gradually leading to differences which may be- 
come so great as to involve an intermediate period of 
change and quiescence, 
There are, however, peculiar difficulties in those cases 
in which, as among the Lepidoptera, the same species 
is mandibulate as a larva, and suctorial as an imago. 
From this point of view Campodea and the Collembola 
(Podura, &c.) are peculiarly interesting. There are among 
insects three principal types of mouth—first, the mandi- 
bulate ; secondly, the suctorial ; and thirdly, that of Cam- 
podeaand the Coilembola generally,in which the mandibles 
and maxillz are retracted, but, though far from strong, 
have some ireedom of motion, and can be used for biting 
and chewing'soft substances. This type is intermediate 
between the other two. Assuming that certain represen- 
tatives of such a type found themselves in circumstances 
* Continued irom p. 109. 
