Houseflies were reared on an artificial medium and tested with different wave-lengths of spectral light obtained from a quartz-mercury arc. The spectrum tested extended from λ3022 Å to λ5780 Å, and the lines were made of approximately equal intensity throughout. In addition, λ5461 Å and λ4078 Å were tested at several other intensities. The comparison standard in all cases was white light, obtained from a tungsten-filament, inside-frosted bulb, and filtered through copper sulphate solution. It was of constant quality, and the intensity was varied by changing the size of the bulb and by varying the distance from the bulb to the copper sulphate filter. The lighted areas to which the flies reacted were 5 by 10 mm. On one of these fell a total intensity of colored light of approximately 10.3 microwatts, on the other a range of intensity of white light of from 0.34 to 36.1 μw.Flies to be tested were removed from the breeding cage ten hours before tests began and were kept in darkness until used. Each fly whose record was used in compiling the final results was caused to make ten trips towards the two test lights, and a record was kept of the choice on each trip.A description and discussion of the four different methods found in the literature for conducting experiments of this type, and for analyzing the results, are included. In the first method, the intensity of the test light of a given wave-length is kept constant, while that of the standard light, usually white, is varied until both are equally attractive.The second method involves testing the colored light against a fixed intensity of white and finding the ratio of insects attracted to color. The intensity of white that will give the same ratio of attractiveness when tested against the standard is then determined.In the third method, the two test lights are made equal in intensity, and their relative efficiency is considered to be directly proportional to the number of insects attracted to each.In the last method, the standard is kept fixed in both quality and intensity, and the intensity of the test color is varied until the two are equal in attractiveness.Application of the first three methods to the same data shows that they give results that vary greatly as the intensity changes. Some show that efficiency increases as the intensity increases, while others show a decrease in efficiency with increasing intensity.If the intensities of all colored lights are equal, the three methods give practically the same qualitative results when applied to the same data. That is, a curve of efficiency is found which has its peak at the same wave-length, whatever method is used. Quantitatively, the results given by the three methods differ, so that no definite ratio of attractiveness can be determined between colors.The data obtained were not amenable to analysis by the fourth method, but published results indicate that this is perhaps the best method for determining the quantitative relation between the stimulative efficiencies of light of different colors.The housefly, M. domestica, is much more strongly stimulated by ultraviolet light of wave-length 3656 Å than by any other part of the spectrum examined. The effect decreases, at first rapidly and then more slowly, as the longer wave-lengths are reached; it also decreases on the short-wave side of the peak. The spectrum available extended only as far as λ3022 Å in the ultraviolet, at which point there was still an appreciable attractiveness, apparently greater than that of either yellow or green.Several problems are suggested that require further investigation.
Constant-Time Complete Visibility for Robots with Lights: The Asynchronous Case
