REPORT ON THE RADIOLARIA. 989 



segments of an incomplete frontal ring (Acrociihus, PI. 93, figs. 15-17). Between the 

 latter and the former there are sometimes developed four tertiary columeUse, as segments 

 of two crossed incomplete diagonal rings {Tympanidium, PL 94, figs. 1, 18). At other 

 times we find six columellte ; here probably the frontal ring is cleft into halves on 

 each side {Tymixiniscus, PL 94, figs. 4-7). Some genera are distinguished by the 

 development of an equatorial ring, or a third horizontal ring (between the upper mitral 

 and the lower basal ring) ; this equatorial ring is either complete [Microcuhus, PL 94, 

 figs. 8-10) or incomplete, developed only laterally {Octotympanum, PL 94, figs. 2, 3). 

 The genus Toxarium (PL 93, figs. 18-20) is distinguished by the development of 

 accessory pairs of bows in the frontal plane, upper or galear bows and lower or thoracal 

 bows, remarkable as beginnings of the accessory joints, which we distinguish in many 

 Spyroidea and Cyrtoidea as" galea and thorax." 



The second subfamily of Tympanida are the Paratympanida, in which the two 

 horizontal rings become closed by lattice-work. The numerous columellse connecting 

 the two rings are here either divergent (Paratympanum, with two unecjual rings, PL 94, 

 fig. 14) or parallel and vertical (Lithotympaniim, with two equal rings, PL 83, fig. 1). 

 The shell of the latter assumes the form of a drum. 



A third small subfamily, Dyst}Tnpanida, is represented by a single genus only, 

 Dystympanuim (PL 94, figs. 15, 16). Here only the upper or mitral ring is closed by 

 lattice-work, whilst the lower or basal ring exhibits a quite simple opening, a large 

 " basal gate " which has probably arisen by loss of the basilar rod of the sagittal ring. 

 These curious forms exhibit a remarkable resemblance to some Monocyrtida, and also to 

 the Dictyocha (Ph.-eodaeia). 



The fourth sul)family, the Eutympanida, are distinguished by the simple wide 

 aperture of the two horizontal rings; the upj)er enclosing a simple " mitral gate," the 

 lower a simple " basal gate." Here probably the two horizontal parts of the sagittal 

 ring (the upper mitral rod and the lower basilar rod) are lost by reduction ; its two 

 vertical parts (the dorsal and ventral rods) only remaining and forming the two paraUel 

 columeUse, which connect the two horizontal rings [Parastephcmus, PL 93, fig. 21). 

 In this group also secondary columeEse are commonly developed between the two rings, 

 alternating with the two primary columellse. Pseudocubus (PL 94, figs. 11, 12), 

 resembling an obelisk, exhibits four divergent columellse ; the two sagittal of these are 

 probably the vertical parts of the primary ring, whilst the two lateral are the halves of an 

 incomplete frontal ring. In Lithocuhus (PL 82, fig. 12) the four columella are parallel 

 and vertical, the two rings being equal ; the shell here assumes the form of a geometrical 

 cube. Prismatium is distinguished by three paraUel columellse ; two of these are j)robably 

 parts of a frontal ring, whilst the third is the remains of a reduced sagittal ring. In 

 Circotympanum (PL 94, fig. 17) and in Lithotympanum (PL 82, fig. 1 1) the number of the 

 columeUse is augmented, there being six to eight or more ; in the former they are divergent 



