252 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 1 37 



a large effect on the birefringence values. Within the time available 

 it was not feasible to dissect and extract sufficient tendons to make a 

 determination on arthropodin from tendons. Instead, as a check on 

 this point, the amplitude of birefringence of cockroach tendons was 

 measured at marked points before and after various treatments. The 

 same values were obtained for fresh tendons in ice-cold water and in 

 absolute ethyl alcohol, and from rewetted tendons after drying. How- 

 ever, an abrupt rise of some 25 to 35 percent followed extraction of 

 the tendon with either hot water or hot lo-percent NaOH solution. 

 This change is qualitatively what would be expected from the separate 

 determinations for chitin, cuticle, and arthropodin given above. The 

 fact that the change was about twice as large as anticipated may or 

 may not be significant.* It may only be a reflection of the fact that 

 soft cockroach cuticle has about twice as much protein as chitin (Den- 

 nell and Malek, 1956; Pfaff, 1952 ; Tsao and Richards, 1952), 



It is obvious from inspection of the above values that the bire- 

 fringence of cuticle can be accounted for by combination of the chitin 

 and arthropodin values only if the arthropodin value is subtracted 

 from the chitin value. But the sign of the birefringence value is posi- 

 tive for both chitin and arthropodin. Therefore, two possibilities 

 exist: (a) the chitin and arthropodin molecules are combined into a 

 complex which itself acts as an optical unit, or (b) the chitin and 

 arthropodin molecules are oriented at right angles to one another 

 (fig. 2 B) . Even with some uncertainty as to the accuracy of the value 

 given for arthropodin, the close agreement of the cuticle value to the 

 chitin minus arthropodin value favors the second possibility. 



It is well known that the optic axis of chitin is the ''b" axis of the 

 crystal lattice, which is the same as the long axis of the chitin chains 

 and the long axis of tendons (Richards, 1951). In other words, a 

 tendon purified to chitin is strongly birefringent when viewed from 

 the side but isotropic when viewed in cross section. But if arthropodin 

 molecules are normally arranged in some regular manner at right 

 angles to the chitin chains, a cross-sectional view of a normal tendon 

 should not be isotropic. Examination of fresh cross sections of cock- 

 roach tendons cut on a freezing microtome and immediately placed 

 in absolute ethanol are birefringent (pi. i, figs. 5-7). Transferring 



* Illness and the imminence of deadline date for submission of this manuscript 

 precluded repeating these determinations on an adequate number of prepara- 

 tions. Of the two determinations made, one showed a 25-percent rise, the other 

 a 33-percent rise, the lower value probably being more accurate than the 

 higher one. 



