LOCALIZED STAGES IN DEVELOPMENT. 163 



is the condensed marginal band of other insects, and the outer, middle, and 

 inner transverse bands, of which the middle lies over the transverse vein 

 across the distal cell. 



The wing in development starts from a minute invagination of cells in the 

 region of the wing spot, which is an area, as shown by Verson, myself, and 

 others, homologous to the spiracular center of other segments. This invag- 

 ination grows inward, and the dorsal portion thickens and grows downward 

 into a tongue-like process either in a pocket of the hypodermis or between the 

 hypodermis and cuticula. These thickenings are found very early, as are also 

 the cell arrangements of the anterior and posterior systems and of the central 

 space. Likewise, before pupation, the interspaces between the veins are fully 

 formed. Later on, in these interspaces in the pupal wings, the enzyme-form- 

 ing cells are found, but of their earlier history we are as yet quite ignorant. 

 The manner of the growth of the wing and the arrangement of the glands, 

 etc., in parallel bands strongly suggest that there exist in the wing fundament 

 of all insects, mother cells, which, by division, give rise to the rows of pigment- 

 enzyme forming cells in the interspaces, and that from these rows they spread 

 in all directions in species of a unicolorous condition, or in special regions 

 only in banded species. It would be difficult to account for these conditions 

 of color development upon any other basis ; and when to this is added all the 

 data of variation, of phylogeny, and of experiment, no other hypothesis seems 

 tenable. It is known that the whole process of wing development is similar 

 in all insects, even, as I have shown, in forms which are as diverse as those 

 with incomplete and complete metamorphosis ; and inasmuch as the initial 

 stages in color development are identical in all, it is highly probable that the 

 above hypothesis is correct. With this hypothesis and the data of ontogeny 

 which so strongly support it, and with our knowledge of the existence of 

 color centers upon the body from which the color patterns develop, we have 

 a basis for the working out and logical interpretation of the phenomena of 

 localized stages in growth and of the laws of the evolution of insect colora- 

 tion and the various phenomena connected therewith. 



LOCALIZED STAGES IN ONTOGENY OF COLORATION. 



The existence in the young of ontogenetic stages or conditions simulating 

 adult characters of related species or genera, ancient and modern, has been 

 abundantlv described and studied in plants, and especially in fossils of both 

 plants and animals. The idea that embryonic, larval, and juvenile conditions 

 are the atavistic reappearances of ancestral adult stages, an old idea revived 

 and modified by L. Agassiz and later crystallized by Haeckel into the dictum 

 that "Ontogeny repeats phylogeny," has been greatly misused, especially by 

 paleontologists, although it is in many cases a valuable aid in phylogenetic 

 study. Jackson, Cushman, and others have described in plants and animals 



