36 FieELp CoLtumMBIAN MusrEumM—GeEo toey, Vot. III. 
tions originate at any considerable depth below the surface. The 
concretions are irregular, lumpy forms without approach to any 
regularity or symmetry beyond the fact that the majority of them 
are more or less flattened and many have one flat side. They are 
occasionally penetrated by minute cylindrical holes up to 2 mm. 
in diameter such as would be the case if they had been penetrated by 
rootlets. They are of reddish-brown limonite color rarely approaching 
a hematite-red in places. They are but slightly consolidated and 
may be readily reducéd to their constituent sand grains by pressure of 
the fingers. They do not commonly exceed 5 centimeters in 
any dimension. In composition they are dune sand cemented by a 
small proportion of limonite which does not fill the voids between the 
grains. The limonite is merely a coating on the sand grains. When- 
ever the grains touch their coatings coalesce, thus cementing the 
sands into a’concretion. There is no evidence of any nucleus in any 
of the specimens examined nor is there any determinable concentric 
structure. 
There is no mystery about the origin of these forms beyond the 
determination of which of three or four common agents has been the 
predominant precipitant of the cement. The sand of the dunes in 
which they were found is, like nearly all dune and beach sand, of a 
yellowish-brown color. This color is due to a thin coating of limonite. 
Where the dunes have not been fixed by vegetation, this color is not 
noticeably lighter at the surface than it isin depth. Where a dune is 
fixed by vegetation a light sod often forms over the surface. Under 
this sod the sand is much lighter in color for a depth of a few inches 
than it is at greater depth. Hence it is to be inferred that the 
organic compounds derived from the vegetation have, as is cus- 
tomary, dissolved the iron oxides from that sand which lies immediately 
under the sod. From organic compounds containing iron dissolved in 
the so-called humus acids, the metal is rapidly precipitated by any 
one of several agents, the more common of which are spontaneous 
changes in the organic solvent, bacterial action, oxidation and hydroly- 
sis. The hydrated ferric oxide precipitated is deposited by preference 
as a film upon the surface of the sand grains and by spontaneous 
dehydration forms the limonite cement. 
As_ he precipitation has fol’owed so immediately on solution as 
to produce concretions within a few inches of the surface it is pro- 
bable that the precipitating agent is either air in the pores between 
the sand grains, iron-secreting bacteria, or more probably a hydro- 
lization of iron compounds of weak organic acids consequent upon 
