380 RESEARCH IN CHINA. 
carbonaceous matter, and even scattered flakes of muscovite. In other 
respects the ground-mass of this rock resembles that of No. 14. The red 
oolitic bodies average smaller (0.6 to 2mm.), but are more uniform in size 
and shape. In their structural features the two varieties are very similar. 
The nuclei consist of irregular bodies of various kinds. Some show 
by shape and structure that they are bits of fossil shells (Plate LIV, 
Fig. E). Others are pieces of former oolites which had been broken up 
(Plate LIV, Fig. F). This fact alone is enough to prove the primary nature 
of the oolites—a fact which has been questioned in some cases. When 
the concentric bodies were formed the deposits must have been uncon- 
solidated. 
Another fact brought out by these slides is that the nodules have all 
been formed under conditions which allowed some of them to be broken 
after having attained full growth, and to be either recemented or to be 
coated with further shells of material; 7. e., the deposit was loose and 
somewhat mobile, and it became solidified only at a later date, probably 
by the crystallization of the calcareous mud in which the particles were 
embedded. 
In one case a bit of black bitumen forms the center of the corpuscles. 
The central portion consists in others of shapeless masses of tangled Girvan- 
ella tubes (Plate LV, Fig. A), but in this rock the supposed alge are confined 
to the interiors of the corpuscles and are evidently nuclei upon which the 
concentric layers of material were afterwards deposited. Finally, there 
are other nuclei of irregular shape and indefinite structure which are not 
recognized. 
The banded outer portions of the nodules are essentially like those in 
No. 14. The rings consist of a carbonate (calcite?) in exceedingly fine 
particles, associated with more or less red hydrohematite. The bands 
conform to the shape of the nucleus in each case. Here also we see a variable 
development of the radial streaks and blotches of iron oxide, in extreme 
cases almost effacing the concentric bands; although very irregular in 
outline the individual blotches tend to have a spindle-shaped or rather 
wedge-shaped form, the broad end being near the margin of the corpuscle. 
One noteworthy fact is that these radiate features are commonest in the 
spherical nodules and rarest on the flat sides of the more elongate bodies; 
1. e., they are more numerous in the situations where any expansion in the 
interior would produce numerous wedge-shaped fissures. Where the radial 
streaks are most abundant they frequently expand in crossing the darker 
rings in the crust and thus anastomose with adjacent streaks. 
The red color is due to finely divided hydrated hematite. The occur- 
rence of red-banded fragments in the nuclei of younger corpuscles shows that 
