92 The Mechanism of Evolution' in Leptinotaksa 



to 4 or 5 weeks before they are mated for the next generation. This period of 

 feeding and inhibition from breeding gives the stock an opportunity to become 

 fully nourished before they begin the breeding activities, and also serves to 

 eliminate the weaker members of the population, which will uniformly die 

 within a few days of emergence. This procedure insures an almost certain and 

 imiform breeding from strong and well-nourished individuals, with the result 

 that inbreeding can be practiced to a far larger extent than if the more usual 

 custom were followed of mating the individuals as soon as they were ready to 

 breed. Another advantage of this method with the sexes isolated and upon 

 normal fresh growing food is that the stock can often be held for months with- 

 out serious consequence ensuing. It is thus possible to breed part of the popu- 

 lation, hold the remainder in reserve for 3, 6, or even 10 months, and thus be 

 able to check the first matings or to make other combinations not made at the 

 start. This is of immense advantage in the working out of the results of 

 hybridization. 



All of my pedigree work with tropical species has been done in large glass 

 houses, with high walls, giving a large volume of air, in which the temperature 

 was controlled automatically, and arranged to correspond with the average 

 night minimum of the growing season in the natural habitat of the species 

 imdergoing investigation at the time. The maximum at mid-day is of less 

 importance and in the summer months was variable, in the sun often from 

 ^9° C. to 35° C. and in the shade from 26° C. to 30° C, which is comparable with 

 the natural conditions found in nature ia the tropical regions from which the 

 material came. The humidity was controlled either by the amoimt of water 

 available for evaporation and by the water of transpiration set free in the air 

 by the plants in the room, whose evaporating surface amoimted on the average 

 to about 25,000 square feet of leaf -surface, or by special humidity control. The 

 actual water-content of the air remained constant or nearly so, but with the 

 range in the daily temperature there was of course a change in the relative 

 humidity of the air of the room and, therefore, of the evaporation from the 

 surfaces of the organisms contained in the cages. 



The food-supply is an important and vital portion of the project and must 

 be met by growing the food appropriate for the species and maintaining a con- 

 stant supply of it, so that at any time matiags can be made as the animals are 

 ready therefor, or when any portion of an experimental series needs to be 

 analyzed. These are best grown as potted plants, after the usual routine of 

 florists, and must be kept free from infestation by pests or the presence of eggs, 

 larvae, or the adults of any of the different species that are under investigation. 

 I have grown my food-supply in an ordinary glass house that was separated 

 from the room where the animals were kept. The food was kept sterile by 

 frequent fumigations with HON gas and by sprayings with solutions of nicotine. 

 This has given complete protection against the possible introduction of any 

 stray individuals into the cultures. 



Many of the food plants that I have used will last for a long time, a year or 

 more, and will give several crops of foliage in the course of their life, thus 

 materially reducing the number of plants that must be grown in the course of a 

 year to maintain the project. My experiments at Chicago have required in the 

 neighborhood of 10,000 plants per year, and a constant supply of from 2000 to 

 3000 plants must constantly be maintained. 



