IN THE LARV^: OF ANISOPTERID DEAGONFLIES. 135 



Eis (23) was found to yield excellent results. The larva is killed in the 



manner shown above. As soon as it is dead, it is removed to a dissectino-- 



dish and pinned under water with its ventral side uppermost. The dissection 



should not be made in glycerine or cedar-oil, since these highly refractive 



media are almost useless for the study of tracheal capillaries. The abdomen 



of the larva is opened by cutting away the projecting edges of the tergites 



and removing the sternites as a continuous ventral plate. The gill-basket 



can then be readily seen, lying in the posterior part of the abdomen. Next, 



the point of a pair of fine scissors is inserted into the posterior opening of the 



basket, and the latter is cut open longitudinally. This operation everts the 



gills, and the basket lies with its inner surface exposed. It is now easy to 



dissect out a complete holobranch. This should be placed on to a clean glass 



slide and floated out with a little water from a pipette. Having arrano-ed it 



in a suitable position (i. e., so as to obtain a good lateral vieio of it), a clean 



cover-glass should be allowed to descend lightly upon it. The gills are thus 



flattened out without being crushed or damaged, and the position of the 



tracheal capillaries is not disturbed. The gills may now be examined under 



a low or moderate power of the microscope, and ii suitable portion selected 



for photography. 



It is very necessary that the dissection, examination, and photographino- of 

 the gill should be completed within an hour or so of the death of the larva, 

 because the air soon afterwards passes out of the capillaries^ which then 

 become invisible. There is no known method by which the air can be 

 retained in these capillaries, so that permament preparations can be obtained • 

 and there is certainly no method that could yield results comparable to those 

 obtainable by the method giveUj in which the capillaries stand out as black 

 lines on a clear backo-round. 



The photomicrographic apparatus used was that of Reichert,, Vienna, 

 arranged in the vertical position. The photographs should be taken by 

 transmitted artificial light (incandescent gas) on slow plates. With Ilford 

 "Process^' plates, which give excellent results^ the exposure varies from 

 fifteen seconds to four minutes, according to the magnification and the 

 aperture of the iris diaphragm. In order to cover the plate it is best to use 

 a No. 4 eyepiece. The most suitable objectives I found to be Nos. 1, 3 6b 

 and 8 a, giving magnifications of 30, 60, 320, and 560 diameters respectively 

 with the closed tube. With the last-named objective a magnification of 

 725 diameters is obtainable by using the full length of the tube, and this can 

 be extended to 850 diameters by lengthening the bellows of the camera. 

 This last magnification is sufficient to show the complete structure of the 

 smallest papillae. One of the fine capillaries in a papilla of Austrocordulia 

 refracta, Till., photographed at 850 diameters magnification, was found to 

 measure barely O'l mm. in diameter. Its actual diameter in cross-section 

 therefore, would be about 0*12 /a. 



10* 



