DEVELOPMENT OF LIMULUS POLYPHEMUS. 195 



attains to a length of nearly three feet ; the Carboniferous types were small and essentially 

 larval, while Cyclus was on the whole emljryonic in appearance. Thus the lowest mem- 

 bers of one group may he the largest, and in a closely allied group, the lowest types be the 

 smallest. In the Decapoda, the largest crabs are found in the highest families, Avhile the 

 lowest Anomoura are the smallest, the lobster many times exceeding Alpheus and Penseus. 

 The lowest Copepoda are the smallest, the Argulus being gigantic compared with the lower 

 forms. So Evadne, etc., the lowest Cladocera, are the smallest, Sida and Daphnia greatly 

 exceeding the former in size. 



In reply to the second objection, i. e., that the young of the three oldest groups of 

 Branchiopoda (Trilobites, Nebaliadoe and Merostoraata) present no relation to a nauplius-like 

 form, I have endeavored to show that the Limulus passes through a nauplius condition soon 

 after the primitive disc appears. The growth of the embryo is at this period remarkably 

 accelerated, and the transitory nauplius state is rapidly effaced, somewhat as Dohrn has 

 shown to be the case in the Decapoda. The third objection falls to the ground if my 

 opinion that the germ at first represents the nauplius condition should be proved. Limulus 

 and the Merostomata as a group, do not, then, offer so remarkable an exception to the other 

 Crustacea as would at first appear, and it should be remembered that the Trilobites in all 

 probability were developed in the same manner as Limulus. 



The dotted lines in the diagram indicate the probable line of development of the 

 Entomostracous Crustacea. We have to assume from the facts stated above, the existence 

 of an Archicaris,^ a hypothetical animal, resembling a larval Copepod, with a rounded, 

 ovate body, not differentiated into head, thorax and abdomen, and with one or two pairs of 

 swimming appendages representing the most anterior cephalic limbs of the Copepoda. 

 This animal we assume to have existed at some time previous to the Silurian age, perhaps 

 towards the close of the Lawrentian period. At certain intervals between the appearance 

 of this Archicaris and the larval Silurian Agnostus, must have appeared a few interme- 

 diate forms of family, or perhaps subordinal, rank, whose offspring partook of the Agnostus 

 form. The transition fi-om a somewhat spherical germ to a flattened Agnostus, may, on the 

 other hand, have occurred, say, in a few thousand generations ; the transition from a 

 spherical germ to a flattened larval form, with a distinct head and abdomen, in geological 

 time, being exactly paralleled by the mode of evolution of the individual Limulus from a 

 spherical, nauplius-like germ, as seen in PI. iii, fig. 13. to a more or less flattened, hemi- 

 spherical subzoean embryo, such as is represented in fig. 19, of PI. iv. We shall have to as- 

 sume, then, the existence of a second main ancestral form uniting the Trilobites and also the 

 Xiphosura (and probably the whole order Merostomata) with the Archicaris. Now 

 Cyclus is the most embryonic of the Merostomata we are acquainted with. The link 

 may have been related in form to this genus, and hence I would propose the term Proto- 

 cycliis for this hypothetical animal, which must if at all have flourished in the Lawren- 

 tian period. In the diagram this Protocyclus may be located at the junction of the dotted 

 lines leading from the Trilobites and Merostomata to the Archicaris. These Lawrentian 

 forms were in all probability not free swimming types like the nauplian young of the 

 Crustacea now living, but very likely crawled over the bottom, in the mud, respiring through 

 their skin like the modern nauplius. They probably adhered to a burrowing life, like the 



^ See Haeckel's Natiirliche Schopfungsgescbichte, 1870, p. 488. 



