214 THE FLORIDA ANTHROPOLOGIST

2006 VOL. 59(3-4)

 

    

as Katee
NT SoZ
LP Sata

Figure 3. Scanning electron microscope images of
Miami Circle pumice artifacts: a) highly elongate or
“tubular” vesicles; b) variable vesicle size associated
with coarsely vesicular pumice; c) euhedral biotite
phenocrysts within vesicular pumice glass; d) high
magnification of a finely vesicular pumice sample; e)
Low magnification of the same sample as in Figure 3d.
A scale bar is provided in the lower right-hand portion
of each image.

surface texture, produced by large (3-5 mm) vesicles (Figure
2d). Vesicles within individual samples are often highly
elongate or tubular in shape (Figure 3a). Other samples,
particularly those with coarser vesicles, exhibit more
equidimensional form (Figure 3b).

Fine- (1 mm) to coarse-grained (5 mm) mineral pheoncrysts
are present in most samples. Large plagioclase pheoncrysts
were noted in several samples (Table 2; Figure 2b). Biotite
phenocrysts (Figure 3c) are present in all samples, however,
their overall abundance is less than two percent. Samples that
have an intermediate composition (andesite) contain relatively
abundant (10-15 %) mafic phenocrysts that include biotite,
amphibole, and altered pyroxene (?) (Figure 2c). Nearly all the
pumice is moderately magnetic. This indicates the presences of
small grains of magnetite within the sample (Table 1).

Scanning electron microscopy imaging of pumice clearly
indicates the two textural types of vesicle shapes. Approxi-
mately fifty percent of the analyzed samples have highly
elongated, “tubular,” vesicles (Figure 3a). The remaining
samples have equidimensional vesicles the may vary in size
(Figure 3 b,d, e). Biotite phenocrysts (Figure 3c) are present as
euhedral grains in most samples.

Modal thin section analyses of a selected set of pumice
samples are reported in Table 2. All the analyzed samples have
a relatively high volume of void space, ranging from 59 to 78
volume percent. Using the appropriate composition for the
samples (Figure 5), the bulk densities for these samples are
calculated to range from 0.40 to 0.70 gm/cm* (Table 2). One

sample has a calculated density greater than 1.0, but the
samples does float in water.

The glass of most pumice samples is normally colorless and
clear. The glassy vesicle walls are very thin, averaging less
than 1-10 yum thick. (Figure 3d). Large, aggregate grains of
plagioclase are common in many samples. Biotite (Figure 3c)
is common in both rhyolite and andesite samples, while
hornblende and cloudy, medium-green aggregates or
glomorocrysts of altered pyroxene (?) or hornblende are present
in low silica rhyolite and andesite. Both biotite and hornblende
are clear and do not exhibit evidence of alteration on the rims
of grains. Mafic samples have either a scoriaceous texture or
have flattened vesicles. The mafic sample from the Custom
House site (8DA1064) is characterized by large, evenly
distributed vesicles that are surrounded by a dark-brown, clear
glass. In addition to glass, the Custom House samples contains
coarse-grained, light brown, subhedral clinopyroxene pheno-
crysts and very fine-grained, euhedral plagioclase. The mafic
pumice sample from the Wynnhaven Beach site (80K239) is
purely vitric, and is characterized by 2-5 mm bands with
alternating dark-green, medium-green and brownish-purple
coloration .

The observed mineral assemblage for most felsic samples is
relatively simple, consisting of plagioclase + biotite + amphi-
bole + quartz + magnetite (Table 2). This mineral assemblage,
along with the absence of mineral phases associated with
peralkaline felsic igneous rocks and the relatively high silica
content (see next section) of the samples are consistent with
classification of a major portion of the samples as being calc-
alkaline rhyolites.

Index of Refraction - Pumice Glass

The optical index of refraction of natural glass is a function
of bulk chemical composition and the post-cooling hydration.
Geologists have used fused rock powder as a means for rapid
categorization of the approximate silica content of fine-grained
volcanic rocks (McKee 1968). While geologists have noted a
high correlation between the SiO, content of natural glasses and
the index of refraction (Tilley 1922), direct use of index of
refraction for the precise determination of silica content is not
possible due to variation in index of refraction by factors other
than SiO, content. These factors include variation of index of
refraction with varying iron and alkali contents for glass having
the same SiO, concentration and post-cooling hydration of
glass, which can produce a higher index of refraction relative to
the same material in an unhydrated state (Ross and Smith
1955). However, the index of refraction method can be readily
used as an aid in “fingerprinting” the provenance of natural
glass samples and in comparing the similarity of material
between different archeological sites. This approach, combined
with geochemical information, has been used as an aid in
obsidian provenance studies. The index of refraction of a
volcanic glass also can be used to test for chemical alteration
and hydration of material following eruption. Departure of a
predicted SiO, content, derived from index of refraction
measurements, relative to chemically determined values, is an
indication of hydration alteration. Index of refraction measure-
ments on Florida pumice can be used for both provenance