FIELD. — INDIVIDUAL VARIATION IN THE LEPIDOPTERA. 391 



The three dimensions defined above were measured with a pair of 

 sharp-pointed dividers with screw adjustment, and a scale graduated in 

 millimetres. To eliminate error, precautions were taken which will 

 shortly be described. 



The measurement of the margin required special apparatus ; and after 

 a number of experiments I succeeded in constructing an inexpensive and 

 eflicient projecting instrument, as follows. A vertical wooden post, about 

 eight inches high, was firmly fixed to one end of a solid base-block. On 

 the top of the post was placed a binding-screw, which held a horizontal 

 brass rod, to the extremity of which was fastened a miniature incandes- 

 cent electric lamp of two candle-power. The electrical connections were 

 formed respectively by the carrying-rod and a flexible spiral of fine wire, 

 the latter also leading to a binding-screw. A one-point switch, situated 

 upon the base-block, served to open and close the circuit. To the lower 

 portion of the post, and projecting horizontally beyond the end of the 

 base-block, was fixed an arm carrying still another binding-screw, hold- 

 ing a vertical steel rod, in such a way as to afford a considerable range 

 of adjustment up and down. To the upper extremity of this rod was 

 fastened a firm "stage" of cork, in which the pin bearing a specimen 

 might be inserted. A graduated scale upon the post permitted an 

 approximate determination of adjustment in height. 



The function of this instrument was to pi'oject an enlarged shadow of 

 the wing upon a sheet of paper placed upon the table, so that the outline 

 might be traced with a pencil. When connected with three cells of a 

 " Samson " open-circuit battery, in a darkened room, it did its work ad- 

 mirably. The shadow produced was sharp and clear, and could be 

 drawn quite as accurately as tlie image formed by an Abbe camera. 



Of course it was not easy to gauge the enlargement (and conse- 

 quent distortion) of the image accurately ; but this was unneccessary. 

 The error being almost exactly the same for the sinuous line and its 

 "chord," (as I have termed the straight line b-d., Figure 1, drawn be- 

 tween its extremities,) it was only necessary to find the ratio between 

 the two, — the actual length of the " chord " being afterward measured 

 directly from the specimen, with the aid of the dividers already referred 

 to. The projected " chord '' was measured with a metric ruler, and the 

 sinuous line with an " Entfernungsmesser," or map-measurer.* 



* The map-measurer is an instrument provided with a small wheel, which is 

 made to roll along the line to be measured, and a recording dial, registering in 

 centiiiK'tres the distance covered. Instruments of this sort are small and rather 

 cheap, and may usually be obtained from dealers in draughtsmen's supplies. 



