THE DEVELOPMENT OF ORGANISMS 791 



original phase in the primitive insect development ; it is of secondary 

 origin, yet with new emphasis. It thus expresses an interpolation, 

 a new chapter, adaptive to certain circumstances, and serving 

 certain purposes. 



Two ideas help us to clearness. The first is that any chapter of a 

 life-history, any arc on life's trajectory, may be in the course of 

 generations lengthened out or shortened down, and a larva repre- 

 sents a lengthening out of a youthful phase. The second idea is 

 that the larva is not the interjection or interpolation of a quite novel 

 form, for that would be magical. The larva is a punctuating off of 

 a stretch in the life-history, with an expression and activation of 

 certain features that would in direct development remain in the 

 background, without being called into active exercise. Furthermore, 

 these larval characters are often adaptive to particular exigencies 

 or difficulties of the creature's life. 



Out of the egg-case (the "mermaid's purse") of a skate, there 

 emerges, after a long, slow development, a fully formed miniature 

 skate, able to fend for itself. There is no larval phase. Out of the 

 floating eggs of the fishing-frog or angler-fish there emerges a quaint 

 long-tasselled larva, suited for life in the open waters, which it 

 leaves for the floor of the sea as it gradually changes into the adult 

 form. The newly-hatched young (Zosea) of a shore-crab is superficially 

 very unlike its parent ; it is so delicate that it could not survive for a 

 day in the rough-and-tumble life of the upper littoral zone; it is 

 suited for the safer open-sea conditions. It feeds, grows and moults, 

 and becomes a "Metazosea". As this second larva assumes by meta- 

 morphosis the familiar adult characters, it sinks, as a Megalopa, 

 to the sea-floor, when not much bigger than a split-pea; thence it 

 gradually creeps up the slope into shallower water, becoming a 

 miniature shore-crab. 



Since every highly-evolved animal does in some measure recapitu- 

 late, in its individual development, the history of its race — ontogeny 

 repeating phylogeny — zoologists used often to regard larvae as 

 rehabilitations of long-lost ancestral types; and this sometimes 

 rightly. In many cases, however, this interpretation is not tenable save 

 in a very general way. It is highly improbable that the caterpillar 

 is at all like any ancestral form of insect; for it is not proved that 

 its attractive resemblance to Peripatus, an annectent type between 

 ringed worms and insects, is much more than of superficial conver- 

 gence. No one could suppose for a moment that the flattened, trans- 

 parent "glass-crab", the Phyllosome larva of the rock-lobster 

 (Palinums) , is at all suggestive of the ancestry of this magnificent 

 crustacean. Yet we must not swing to the opposite extreme; the 

 surface larvae of the floor-frequenting flat fishes are nearer the 

 ordinary and ancestral fish-type ; and larval ascidians certainly give 

 us more than a hint of the vertebrate, pedigree of these strange 



