:-l-l THE CANADIAN ENTOMOLOGIST. 



the constant term, one making the temperature two or three degrees lower 

 than the other. Edes does not say that the temperatures were read from 

 a standard tested thermometer, and the discrepancy noted was just such 

 as might be expected from the use of incorrect thermometers. 



II. — Statement of Problem and Method. 



To test the formulas of Dolbear and Bessey, I counted the chirps of 

 several crickets, and from them computed the temperatures. Finding that 

 the computed temperatures were sometimes considerably in error, I 

 undertook a series of observations to determine (i) whether the discrepancy 

 was to be found in any very large proportion of the crickets, and if so, (2) 

 to discover, if possible, the cause of the discrepancy. 



These observations were made on CEcanthus iiiveiis in Ann Arbor, 

 Mich.,. in the level country near New Carlisle, Ohio, and on the hillsides 

 of Ithaca, N. Y.,in the summers of 1905 and igo6. At first all observa- 

 tions were made with the insect in view. Lantern in hand, I cautiously 

 approached the point in the bushes from which the sound issued until the 

 chirping insect was discovered and the chirps found to coincide with the 

 motion of the wings. The crickets were not at all disconcerted by this 

 procedure, and often permitted me to approach within two feet. 



After having observed some 75 specimens by means of a lantern, I 

 found that chirping crickets were rarely closer together than three or four 

 feet, and hence that an individual could easily be picked out by sound. 

 Thereafter I relied uj^on sound alone for the determination of rates. Of 

 the observations made with a lantern, the only one. recorded in the 

 following pages is the one mentioned in the discussion as having chirped 

 2,228 times in succession. An ordinary watch was used in timing, and 

 to minimize the error, each count was continued through several mmutes. 

 Temperature readings were taken within a few minutes of making the 

 count, and all readings were made from the same thermometer. This 

 thermometer was afterward compared with a tested thermometer, by 

 immersing them simultaneously in vessels of water at various temperatures, 

 and was found to have a constant error of about half a degree within 

 tlie range of temperatures recorded below. Corrections have accordingly 

 been made in these readings before entering them in the tables. The 

 thermometer read to degrees, and fractions of degrees were estimated. 

 The length of one degree on the scale was such that for an eye trained in 

 estimating fractions of lengths the error should in any case have been less 

 than one-tenth of a degree. 



I am indebted to Prof. Jacob Reighard for much assistance in 

 discussing my data and in preparing this paper for publication, 



