TRANSACTIONS OF SECTION D. 1091 
hive-bees, humble-bees, wasps, and other insects, where such a difference occurs, but 
where the young are directly fed, it is, of course, obvious that the quantity can be 
proportioned to the appetite of the grub. But in insects with the habits of Zwmenes 
and Ammophila, the case is different, because the food is stored up once for all. 
Now it is evident that if a female grub was supplied with only food enough for a 
male, she would starve to death ; while if a male grub were given enough for a 
female, it would have too much. No such waste, however, occurs. In some 
mysterious manner the mother knows whether the egg will produce a male or 
female grub, and apportions the quantity of food accordingly. She does not change 
the species or size of her prey, but if the egg is male, she supplies five; if female, 
ten victims. Does she count’ Certainly this seems very like a commencement of 
arithmetic. At the same time it would be very desirable to have additional evidence 
how far the number is really constant. 
Considering how much has been written on instinct, it seems surprising that so 
little attention has been directed to this part of the subject. One would fancy 
that there ought to be no great difficulty in determining how far an animal could 
count, and whether, for instance, it could realise some very simple sum, such as 
that two and two make four. 
But when we come to consider how this is to be done, the problem ceases to 
appear so simple. We tried our dogs by putting a piece of bread before them, and 
preventing them from touching it until we had counted seven. To prevent our- 
selves from unintentionally giving any indication, we used a metronome (the in- 
strument used forgiving time when practising the pianoforte), and to make the 
beats more evident we attached a slender rod to the pendulum. It certainly seemed 
as if our dogs knew when the moment of permission had arrived ; but their move- 
ment of taking the bread was scarcely so definite as to place the matter beyond a 
doubt. Moreover dogs are so very quick in seizing any indication given them, even 
unintentionally, that, on the whole, the attempt was not satisfactory to my mind. 
I was the more discouraged from continuing the experiment in this manner by 
an account Mr. Huggins gave me of a very intelligent dog belonging to him. 
Cards were placed on the ground, numbered from 1 to 10; and a question being 
6+2-3 
then asked: the square root of 9 or 16; or such a sum as , Mr. Huggins 
pointed consecutively to the cards, and the dog backed when he came to the right 
one. Now Mr. Huggins did not consciously give the dog any sign, yet so quick 
was the dog in seizing the slightest indication, that he was able to give the correct 
answer. This observation seems to me of great interest in connection with the so- 
called ‘ thought-reading.’ No one, I suppose, will imagine that there was in this 
case any ‘ thought-reading’ in the sense in which this word is used by Mr. Bishop 
and others. Evidently the dog seized upon the slight indications unintentionally 
given by Mr. Huggins. I have brought this question before the Section in hope of 
inducing others with more leisure and opportunity to carry on similar observations, 
which I cannot but think must lead to interesting results. 
3. On the Development of the Food-fishes at the St. Andrews Marine 
Laboratory. By Epwarp E. Prince. 
After referring to the literature of the subject and the incomplete state of our 
lmowledge of the embryology of the osseous fishes of British seas, the author pro- 
ceeded to give details of the deposition of the ova in certain species. Of about 
twenty forms, deep-sea and littoral, studied in the St. Andrews Marine Laboratory, 
attention was specially directed, chiefly on account of their economic importance, 
to the six following“ species, viz., Gadus merlangus, G. morrhua, G. e@glefinus, 
Trigla gurnardus, Pleitronectes flesus, and P. limanda. Amongst these species, 
the ova of which are pelagic, differences in the manner and duration of deposition 
probably obtain. Thus the extrusion of ova in the Pleuronectidz would appear to 
be more rapid and continuous than in Trigla gurnardus and the gadoids, in which 
the act of spawning is, it would seem, intermittent and prolonged. The examina~- 
4a2 
