He conceived that progressive evolution might take place in two different ways. "Descendants ... reach a new goal, either by deviating sooner or later whilst still on the way towards the form of their parents, or by passing along this course without deviation, but then instead of standing still advancing still farther" (Eng. trans., p. 111). In the former case the developmental history of descendants agrees with that of the ancestors only up to a certain point and then diverges. "In the second case the entire development of the progenitors is also passed through by the descendants, and, therefore, so far as the production of a species depends upon this second mode of progress, the historical development of the species will be mirrored in its developmental history" (p. 112). Of course the recapitulation of ancestral history will be neither literal nor extended. "The historical record preserved in developmental history is gradually _effaced_ as the development strikes into a constantly straighter course from the egg to the perfect animal, and it is frequently _sophisticated_ by the struggle for existence which the free-living larvae have to undergo" (p. 114). It follows that "the primitive history of a species will be preserved in its developmental history the more perfectly the longer the series of young stages through which it passes by uniform steps; and the more truly, the less the mode of life of the young departs from that of the adults, and the less the peculiarities of the individual young states can be conceived as transferred back from later ones in previous periods of life, or as independently acquired" (p. 121). Applying these principles to Crustacea, he concluded that the shrimp _Peneus_ with its long direct development gave the best and truest picture of the ancestral history of the Malacostraca, and that accordingly the nauplius and the zoaea larvae represented important ancestral stages. He conceived it possible so to link up the various larval forms of Crustacea as to weave a picture of the primeval history of the class, and he made a plucky attempt to work out the phylogeny of the various groups. The thought that development repeats evolution was already implicit in the first edition of the _Origin_, but the credit for the first clear and detailed exposition of it belongs to F. Mueller. In much the same form as it was propounded by Mueller it was adopted by Haeckel, and made the corner-stone of his evolutionary embryology. Haeckel gave it more precise and more technical formulation, but added nothing essentially new to the idea. It is convenient to use his term for it--the biogenetic law (_Biogenetische Grundgesetz_)--to distinguish it from the laws of Meckel-Serres and von Baer, with which it is so often confused.