360 ROBERT TRACY JACKSON ON ECHINI. 



are narrower than on the exterior, and demi-plates, opposite horizontal interambulacral sutures 

 are higher and laterally fan-shaped (compare text-fig. 244, p. 338). The pore-pairs on the inner 

 aspect lie near the outer border of the occluded plates as they do on the exterior, but they lie 

 near the inner border of the demi-plates, and in all isolated plates the pore-pairs are approxi- 

 mately in the middle of each plate, instead of near the outer margin of the plate, as they are in 

 all plates on the exterior (Plate 61, figs. 5-9). Ambulacral plates bear small secondary tubercles 

 and spines similar to those of the interambulacra (Plate 61, fig. 8), (pp. 54, 59). 



Such being the structure of the ambulacrum at the mid-zone, the ventral developmental 

 characters are of great interest. Near the peristomal border in Melonechinus there are typi- 

 cally four columns of ambulacral plates (Plate 56, figs. 2, 3), narrow demi- and wider occluded 

 plates; this is the character of plates at the mid-zone in Lovenechinus (Plate 42, fig. 2), and 

 as a third stage in the development of Oligoporus (Plate 50, fig. 8). Passing dorsally in Melon- 

 echinus (Plate 56, fig. 3), soon scattered isolated plates appear in the middle of each half-area; 

 this is like the character at the mid-zone in Oligoporus (Plate 50, fig. 7). Again, passing dor- 

 sally in Melonechinus, the isolated plates become more frequent, so as to come in vertical con- 

 tact, and thus a new column is attained and the essential generic character acquired. If the 

 species is to have only six columns of plates at the mid-zone, the ventral development goes 

 no further, but in others additional isolated columns appear with dorsal growth until the full 

 complement of the species in hand is attained. From this it follows that species with many 

 columns of isolated plates pass through stages with fewer columns, which stages are directly 

 comparable to the condition at the mid-zone of lower species in the genus, as briefly shown in 

 the diagram (text-fig. 237, p. 231). 



While this is the typical character of the development of the ambulacrum in Melonechiims, 

 as I have traced it in several species and many specimens, in one specimen of Melonechinus 

 multiporus a still simpler condition occurs. In this choice individual (Plate 57, figs. 1,3; text- 

 fig. 245, p. 382), in area B, ventrally there are a few primary plates, which pass directly across 

 the half-areas, and, while not typical as a ventral developing character in Melonechinus, these 

 plates express a reversionary radial variation to the character typical of adult Palaeechinus, 

 the lowest genus of the family, and also to a condition that is typical of the ventral border as 

 a developing stage in Maccoj'a, Lovenechinus, and Oligoporus. And yet some people maintain 

 that the recapitulation theory is a myth, and stages in development do not exist ! 



While the ventral developing stages are the most clear, yet dorsally the ambulacrum in 

 Melonechinus presents a series of localized stages in the young last added plates. Next to 

 the ocular are found few, even as few as two plates (Plate 53, fig. 1; Plate 56, fig. 6), and, when 

 there are two, the condition represents two columns as a stage, which is simpler than the 

 usual ventral stage with four columns and is comparable to the typical condition of Palaeechinus. 

 Passing ventrally to the next older plates, we find four columns (Plate 52, fig. 2), comparable 



