Sept., 1903.] Needham & Anthony: The Thorax of Odonata. 119 



senior author bethought himself that this skewness might be measured, 

 and devised as an instrument for the purpose the goniometer shown in 

 PI. VIII, Fig. I. This was constructed with little trouble out of a 

 discarded box top, about 100 mm. square, a small brass protractor 

 scale, a bicycle spoke, and a piece of brass about 25 mm. stjuare. The 

 brass was first drilled through the center and reamed out so that the 

 head of the bicycle spoke would fit it neatly and rotate in it smoothly. 

 Then the corners of the brass were drilled to receive screws. Then, 

 with the spoke in place, its head flush with the surface of the brass, 

 the latter was screwed fast to the under side of the wooden base, 

 nearer the hinge edge of the cover, from which the side strip that was 

 underneath had been removed. Thus the spoke was securely held 

 by its head while free to rotate in the brass. Then the spoke was 

 bent twice at right angles in an elongate U with unequal arms, the 

 first bend perpendicular to, the second parallel to the surface of the 

 board, the two arms being strictly parallel and far enough apart to 

 allow the placing of the body of the largest dragonfly between the 

 upper arm and the wooden base (PI. VIII, Fig. i). Then the pro- 

 tractor scale was glued to the wooden base in such position that its 

 center of curvature was exactly over the center of the pivot below the 

 base. The longer upper arm of the U then crossed the center in any 

 position of rotation, and its end crossing the scale served as an indi- 

 cator. 



To use this goniometer a dragonfly with wings folded back to back 

 was laid on a broad glass slip (this merely for convenience in moving 

 the specimen) and brought to rest with its predetermined base line of 

 angle measurement coinciding with the base line of the protractor 

 scale. Then the index arm above was moved parallel with the suture 

 forming the other limb of the angle to be measured. Then the angle 

 was read by sighting along the edge of the indicator, keeping the ex- 

 act center and the degree to be read in alignment. Thus the three 

 successive operations — the placing of the specimen, the adjustment of 

 the indicator and the reading of the scale — were done independently 

 and in the order stated. This made for accuracy, but there were both 

 mechanical and anatomical reasons why great accuracy was unattainable. 



I. Mechanical. — The base line was too short. It was impossible 

 to go beyond the confines of the combined meso- and metathorax and 

 have fixed points, owing to the flexibility of the articulations with i)ro- 

 thorax and with abdomen. To settle upon two points that should 



