IEPISMA SCALES. 167 



vented from moving by a little paste at each corner. 

 No. 3 may then be taken as an exaggerated section of 

 the various parts. A B is the glass slide, with a scale, 

 C, closely adherent to it, and D the thin glass-cover. 

 If a very small drop of water be placed at the edge of 

 the thin glass, it will run under by capillary attraction ; 

 but when it reaches the scale, C, it will run first 

 between it and the glass slide A B, because the attrac- 

 tion there will be greater, and consequently the mark- 

 ings on that side of the scale which is in contact with 

 the slide will be obliterated, while those on the other 

 side will, for some time at least, remain unaltered : 

 when such is the case, the strongly marked vertical 

 lines disappear, and the radiating ones become con- 

 tinuous. (See No. 1, the lower left-hand portion.) 

 To try the same experiment with the other, or inner 

 surface of the scales, it is only requisite to transfer 

 them, by pressing the first piece of glass, by which 

 they were taken from the insect, upon another piece, 

 and then the same process as before may be repeated 

 with the scales that have adhered to the second slide ; 

 the radiating lines will now disappear, and the vertical 

 ones become continuous. (See No. 2, left portion.) 

 These results, therefore, show that the interrupted 

 appearance is produced by two sets of uninterrupted 

 lines on different surfaces, the lines in each instance 

 being caused by corrugations or folds on the external 

 surfaces of the scales. Nos. 1 and 2 are parts of a 

 camera lucida drawing of a scale which happened to 

 have the opposite surfaces obliterated in different 

 parts. No. 4 shows parts of a small scale in a dry and 

 natural state ; at the upper part the interrupted appear- 

 ance is not much unlike that seen at the sides of the 

 larger scales, but lower down, where lines of equal 

 strength cross nearly at right angles, the lines are 

 entirely lost in a series of dots, and exactly the same 

 appearance is shown in No. 5 to be produced by two 

 scales at a part where they overlie each other, although 

 each one separately shows only parallel vertical lines." 

 Another very characteristic fallacy resulting from 

 configuration is furnished in the supposed tubular 



