carapace serve to place Burgessia in the Branchiopoda under the order Notostraca.
Among the anostracans Opabinia regalis, in its elongate many-segmented body, phyllopod-like swimming exopodites and insignificant or rudimentary ambulatory endopodites, small head, and slender body, is very suggestive of an annelidan ancestor.
These comparisons raise the question as to the relations of the Branchiopoda, Leptostraca (representing the Malacostraca), Trilobita, and Merostomata. With the data afforded by the Burgess shale fauna the inter-relationship of the four so-called subclasses is found to be very intimate.
In Opabinia (pl. 27, fig. 6, and pl. 28, fig. 1) and Leanchoilia (pl. 31, fig. 6) the typical branchiopod is clearly present.
In Waptia (pl. 27, figs. 4 and 5) the Leptostraca is very near at hand as developed in Hymenocaris (pl. 31, figs. 1 and 2).
In Marrella (pls. 25 and 26) the trilobite is foreshadowed, and Nathorstia (pl. 28, fig. 2) is a generalized trilobite as the trilobite appears to be a specialized branchiopod, adapted largely for creeping on the bottom. The trilobite gives some conception of a possible form between the Branchiopoda and the Aglaspidæ of the Merostomata.
Such forms as Habelia (pl. 29, fig. 6), Molaria (pl. 29, figs. 1-5), and Emeraldella (pl. 30, fig. 2) serve to fill in the gap between the Branchiopoda and the Merostomata as represented by Sidneyia[1] and later the Eurypterida. Sidneyia is now known to have a pair of jointed biramous appendages on each of the anterior 9 segments of the body. The inner division or endopodite is a jointed leg adapted for creeping close to the bottom and the outer branch is a lamellated branchial lobe (see Smithsonian Misc. Coll., vol. 57, No. 2, 1911, pl. 6, fig. 3, and pl. 7, fig. 1; and text fig. 10 of this paper, p. 206).
In the following diagram the attempt is made to show the relations of Cambrian crustaceans to a theoretical ancestral stock which for convenience is correlated with the Apodidæ. From this stock it is assumed that the Branchiopoda came, and from the Branchiopoda stock three distinct branches were developed prior to or during Cambrian time. Of these the one of greatest interest in the present connection is that on the right of the diagram. In this line of descent it is assumed that the Trilobita are directly descendent from the Branchiopoda and forms grouped under the order Aglaspina derived from the Trilobita. The order Limulava is considered as being inter-
