RESFIBATOKY OECUNS IN AKANBJ]. 
55 
through the tracheal and pulmonary proliferations respectively 
of one and the same embryo from a series of sections similar 
to those from which fig. 27 has been reconstructed and re- 
presenting the same stage. It is to be noticed that the tracheal 
proliferation {tr. proZ. ) is solid throughout, while that of 
the pulmonary sac (pulm. *•.) is provided with a considerable 
cavity, though this is, of course, not a fundamental difference 
but is to be considered rather as due to a mere difference in 
the rapidity of growth. In other respects both proliferations 
closely resemble each other: in each the incision between it 
and the outer epithelium is deepest on the anterior side and 
dorsally at the apex. The opening of the tracheal sac does 
not extend dorsally beyond section Ko. 10 (fig. 6a) of fig. 27, 
the following five sections (compare figs. 35 and 35a) showing 
no trace of a post-appendicular groove, exactly as in the case 
of the corresponding five sections (Nos. 25-29) of the pul- 
monary appendage. In both cases the dorsal ends of the 
openings represent the latero-dorsal ends of the permanent 
spiracles, the medial ends of which are still unformed. 
Shortly after the stage I have just described the migration 
of the three posterior pairs of abdominal appendages, already 
alluded to on a previous page, commences. This process, which 
may be considered as characteristic of all Dipneumonous 
spiders with the tracheal spiracle near the hind end of the 
body, consists of a double movement, namely, a medio-ventrad 
movement of each of the three pairs of appendages and a 
caudad one caused by the enormous elongation of the ninth 
somite. Near the end of the reversion, as a result of this 
process, these appendages come together in pairs in the median 
line in the posterior half of the abdomen {tr. pL, figs. 41 and 
43). At the same time the tracheal appendages gradually 
sink to the level of the body surface. 
During this period the formation of the tracheal spiracles 
is completed, the lateral ends of the spiracles having already 
been formed at an earlier stage. The unformed median ends 
become approximated by the migration of the appendages 
towards the median line, and subsequently the region of the 
