Content uploaded by Keith H. Ashley
Author content
All content in this area was uploaded by Keith H. Ashley on Jun 27, 2016
Content may be subject to copyright.
The Florida Anthropologist
University of North Florida
Vol. 69 (1) March 2016
Keith Ashley, Vicki Rolland, and Robert L. Thunen
EXCAVATIONS AT THE ARMELLINO SITE (8DU631): THE PROPOSED MOCAMA
VILLAGE AND VISITA OF SARABAY
In 1998 and 1999 the University of North Florida (UNF)
conducted summer eld school excavations at the Armellino
site (8DU631) on Big Talbot Island. This barrier island is
situated between the St. Johns and Nassau rivers in extreme
northeastern Florida. These excavations followed on the heels
of a systematic shovel test survey of the southern third of
the island, performed by UNF in the Spring of 1998 (Ashley
and Thunen 2000, 2008). The Armellino site is located on
the leeward side of the island adjacent to Mud River, which
would have provided an important access point onto the
island from the inland sound (today’s Intracoastal Waterway).
A high density concentration of San Pedro and San Marcos
pottery1, identied as a result of shovel testing, suggested the
Armellino site might represent the location of a documented
Mocama village and mission visita known as Sarabay (Figure
1). Excavation of an 8-x-7 m block (A-B) during the two
eld seasons yielded sixteenth and early seventeenth-century
artifacts and exposed a portion of what appears to be a native
wall trench structure along with more than 50 subsurface
features, most of which seemingly represent postmolds and
pits. This paper presents highlights of the 1998-1999 eld
school excavations, focusing on features and artifacts.
Mocama Archaeological Project
In 2009 the UNF Archaeology Lab initiated the
Mocama Archaeological Project (MAP), a collaborative,
multidisciplinary research program designed to combine
archaeological survey, excavation, and standardized and
specialized analysis; GIS mapping and remote sensing; and
documentary and archival research in Florida, France, and
Spain. MAP is committed to the search for Mocama villages
and associated European colonial communities in order to
reconstruct the sixteenth and seventeenth century social
landscape of northeastern Florida. Beyond locating these
settlements and exploring their physical layouts, this long-term
program is designed to research the social history and culture
of the region’s indigenous population in the face of European
contact, colonization, and missionization. Our goal is to
focus on natives and newcomers as dynamic and interacting
communities rather than as static and isolated entities.
Our initial phase of research concentrated on Mocama
settlements situated on barrier islands north of the St. Johns
River. This began with a detailed reanalysis of the artifacts
recovered from the Armellino site in 1998 and 1999. As
discussed below, this site is considered to represent the
Contact period Mocama village and mission visita of Sarabay,
Next, we undertook a 3-year (2009-2011) archaeological
investigation of the Cedar Point West site (8DU63), which
we believe is the contact settlement and Spanish visita known
to the Spanish as Vera Cruz (Ashley et al. 2010, 2012).
Fieldwork also was conducted at the nearby Cedar Point site
(8DU81), which is the archaeological location of the relocated
Franciscan mission of Santa Cruz de Guadalquini, 1684-
1696 (Ashley et al. 2013a, 2013b). Both archaeological sites
are situated at the southern end of Black Hammock Island
within the National Park Service (NPS) Timucuan Ecological
and Historic Preserve (Russo et al. 1993). A survey also was
conducted to sample a series of small sites, immediately north
of the mission of San Juan del Puerto on Fort George Island,
where Contact and Mission period ceramics were known to
occur (Ashley and Gorman 2011).
Most recently, UNF conducted an archaeological survey
in search of evidence of the sixteenth-century French colony
of La Caroline (or the subsequent Spanish Fort San Mateo).
During the course of eldwork, 725 50-cm2 shovel tests and
one 1-x-2-m unit were excavated. Thirteen archaeological
sites were newly discovered or relocated on National Park
Service (NPS) and private lands in the vicinity of Fort Caroline
National Memorial along the south bank of the St. Johns River
(Ashley et al. 2013). Although a small amount of Contact-era
San Pedro pottery was recovered, no evidence of a sixteenth-
century Mocama Indian village was discovered. Moreover, no
sixteenth-century French or Spanish artifacts relating to the
La Caroline colony (Fort Caroline) or the subsequent Spanish
Fort San Mateo, were found. Thus, questions on the location
of the French and Spanish forts still linger and await future
research.
In summary, based on these projects and the results
of earlier archaeological investigations and documentary
evidence, we have tentatively correlated a series of
archaeological sites in northeastern Florida with their named
historic village counterparts (Figure 2). These are presented
in Table 1 and discussed elsewhere (Ashley 2009:135-138;
2013:161-163, 2014:164-171). For the purposes of this paper,
we will focus on the Armellino site and the village of Sarabay.
Mocama Village of Sarabay
The European invasion of America struck the Mocama
coast of northeastern Florida and southeastern Georgia with
the arrival of French Huguenots, rst in 1562 to explore
and again in 1564 to colonize (Bennett 1975; Lawson 1992;
Ribault 1964; see Ashley 2014 for a recent overview of
Mocama archaeology). With the building of Fort Caroline
near the mouth of the St. Johns River in 1564, this tidewater
The Florida Anthropologist
50 2016 69 (1)
Figure 1. Location of Sarabay (Armellino site) and other post-contact sites in the area.
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 51
Figure 2. Suspected location of historic Mocama villages.
The Florida Anthropologist
52 2016 69 (1)
zone became the focal point of late sixteenth-century French-
Indian relations along the southern Atlantic coastline. During a
brief 15 months among the Mocama, French chroniclers made
only a few vague references to a village known as Serranay
(or Sarrauahi, Saranai). Rene Laudonnière, commander of
the French fort, reported that the village of Serranay was
situated “…about a league and a half from our fort on an
arm of the river” (Bennett 1975:96; Lawson 1992:79). He
reafrmed this general location in a later passage in which
he states “…in the village of Serranay, situated on the other
side of the river and two leagues from the fort…there were
elds in which the maize was ripening and there was a great
deal of it” (Bennett 1975:139; Lawson 1992:116). Moreover,
the vicinity of Serranay was apparently visited “every day”
by French workers “to get clay to make bricks and mortar
for our houses.” The chief of Serrany is also mentioned by
Ribault and Laudonnière on several occasions as being present
in meetings between the French and local Mocama “kings”
(Bennett 1975:156; Lawson 1992:127).
The French colony of La Caroline only lasted a little more
than a year, and by late 1565 it fell to Spanish forces under the
command of Pedro Menéndez de Avilés. By 1587, the Spanish
had established two missions within Mocama territory: San
Juan del Puerto on present-day Fort George Island, Florida and
San Pedro de Mocama on today’s Cumberland Island, Georgia.
In a letter written by Fray Franciso Pareja in 1602, the friar
states “I have been taking part in the conversion and teaching
of these natives for seven years and that during all this time I
have served in this village of San Juan del Puerto called the bar
of San Mateo, instructing and baptizing; both in this village
where I am vicar and in the settlements that are neighbors to
it…” (Pareja 1602). He goes on to list nine villages (visitas)
along his evangelical circuit including Çarabay, the closest
to San Juan at “a fourth a league away.” The visita and later
mission of Santa María de la Sena was located to the north
on Amelia Island, which in 1602 was under the jurisdiction
of Fray Baltasar López at mission San Pedro de Mocama on
Cumberland Island, Georgia (López 1602).
The village of Sarabay (or Çarabay) is not mentioned in
any known Spanish documents after 1602 (Hann 1990:456).
However, the name persists in Spanish records and maps to
refer to an island between the bay or bar of San Mateo (mouth
of the St. Johns River) and the bar of San Pedro (St. Marys
River). In a document known as the González derrotero
(course) of 1609, Andrés González, pilot for 1609 Ecija
expedition to Jacán (Virginia), provides rich detail of the
coastline of northeastern Florida and southeastern Georgia
(Hann 1986). After leaving St. Augustine and passing the bar
of San Mateo and before encountering the bar of San Pedro,
González mentions the bars of Sarabay (Fort George River)
and Sena (Nassau River), which “are a matter of a league
and a half from the bar of San Mateo” and are “not bars of
importance” (Hann 1986:48).
Additional Spanish documentary and cartographic
information clearly identies the barrier island north of San
Juan as Sarabay. Knowing unequivocally that San Juan was
situated on Fort George Island and that Sena is the name used
in Spanish mission documents for Amelia Island, it strongly
suggests that Sarabay is present-day Big Talbot Island, because
no other island of any size existed between them at the time.
The distance provided by Pareja points to the southern end of
Big Talbot as the most likely location for the early mission
visita of Sarabay. Based on the Spanish practice of embedding
missions and visitas within the existing native settlements, it
is probable that early mission settlements were at or very near
their Contact-era counterparts (Ashley 2013:151). If true, then
we should expect to nd evidence of the Contact-era village
(1562-1587) and early mission visita (1587-1605?) of Sarabay
on the same archaeological site or at the least very close by at
the southern end of Big Talbot Island.
Armellino Site (8DU631)
Situated in a maritime hammock along the eastern edge
of Mud River near the southwestern tip of Big Talbot Island,
the Armellino site covers approximately 800 (N-S) by 250 (E-
W) m (Figure 3). The terrain is rather at, rising only a meter
or two above the adjacent saltmarshes. The site lies between
8DU90 to the north and 8DU628 to the south and is bounded
by Houston Road to the east and tidal marshes to the west. The
Houston Cemetery (8DU1549) is located within the south-
central part of 8DU631. Most of the site occurs within the
boundaries of Big Talbot Island State Park, but its approximate
center extends onto private property. A large wetland lies
immediately east of the site. As in early historic times, Big
Talbot continues to be sparsely populated with a few houses
constructed along its southwestern banks.
Previous Archaeology
Site 8DU631 was rst recorded by Nidy (1980:52) during
a walkover of portions of the island focused on roads, trails, and
other areas of surface visibility. The only artifacts recovered at
this time (1976) were two olive jar fragments from a road cut,
leading Nidy to estimate a site size of “probably less than one
acre.” He further suggested that the sherds might be related
to a “Spanish mission station, or visita, located in the general
area” (Nidy 1980:52).
Nidy was not the rst to equate the southern end of Big
Talbot Island with Sarabay. That distinction goes to William
Documented site Archaeological Site
San Juan del Puerto/
Alimacani
San Juan del Puerto (8DU53)
Vera Cruz Cedar Point West (8DU63)
San Mateo (visita)Riverwoods site (8DU11831)
San Pablo Greeneld sites (8DU5544/45)
Sarabay Armellino (8DU631)
Santo Domingo Old Town (8NA9)
Santa María de Sena Harrison Homestead (8NA43)
Table 1. Documented historic Mocama villages and
suspected archaeological sites.
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 53
Figure 3. Shovel testing of southern third of Big Talbot Island, with boundary of the Armellino site (8DU631) shown.
The Florida Anthropologist
54 2016 69 (1)
1998-1999 UNF Field Schools
Following the shovel test survey, UNF eld schools
spent ve weeks in the summer of 1998 and ve more in
1999 testing portions of the Armellino site. In addition to
instructing students in proper eld techniques, the purpose
of the project was to investigate the site’s Contact and early
Mission period components. Initial unit placement was guided
by shovel test results and focused on locations that had yielded
high quantities of San Pedro and/or San Marcos pottery or
possible subsurface features. Each unit was excavated in 10
cm arbitrary levels that were subdivided as needed. All ll was
screened through one-quarter inch mesh, and the quantity of
shell from each excavated provenience was recorded in liters
prior to discard. Soil samples (removed in their entirety) were
taken from most features and selected shell midden contexts
for ne mesh (1/16”) water screening in the lab. Artifacts from
these samples were analyzed, while shell, bone, and charcoal
were sorted and are available for future analysis.
Testing began with the excavation of four 1-x-2 m units.
Units 1-3 were placed near the approximate center of the site,
while Unit 4 (in Block A) was positioned south of the historic
cemetery (Figure 4). Unit 4 expanded upon an earlier shovel
test that partly intersected a subsurface feature (i.e., organic
stain). Additional staining was encountered at the midden-
subsoil interface (~40 cm below surface) of Unit 4, prompting
the excavation of additional 1-x-2 m units. In all, nine units
comprised Block A, which measured 3-x-6 m. The following
summer (1999) Block A was expanded to the north and east
to form Block B, which consisted of 15 1-x-2 m units (one
of which was not excavated due to the presence of a tree).
Together the two contiguous blocks covered roughly 8-x-7 m.
In addition to these block excavations a 1-x-3 unit and a 1 m2
were dug into Shell Midden A, located 20 m to the northeast.
In all, 56.25 m2 were excavated during the two eld seasons
(see Figure 4).
Unit 1 (4251N/1026E)
Unit 1 was placed within a private residential lot near
the approximate center of the site, where shovel testing had
yielded a moderate amount of San Pedro pottery. Six 10 cm
levels were excavated, revealing a thin humic layer of dark
gray (10YR 4/1) sand followed by an approximately 40 cm
thick zone of very dark grayish brown (10YR 3/2) sand above
a sterile grayish brown (2.5Y 5/2) ne sand. Soil from the
last two levels was very wet. Sparse shell and animal bone
were recovered, with the former found in Levels 1-5 and the
latter in Levels 3-5. Pottery consisted of 143 sherds (712.6 g)
larger than 2 cm in size, with the majority (84.6 percent by
count and 87.5 percent by weight) coming from Levels 2-4
(Table 2). More than 50 percent of the assemblage consisted
of St. Marys Cordmarked (A.D. 1250-1450), with San Pedro
making up 11.9 percent of the assemblage by count and 14.4
percent by weight. San Marcos composed just 5.6 percent of
the assemblage by weight and 8.9 percent by weight. Other
pottery types identied in Unit 1 include St. Johns, Swift
Creek, and sand tempered.
Jones, an avocational archaeologist interested in nineteenth-
century plantation sites on the island. Jones (1988:9)
suggested that the visita might exist “[s]outh of the Houston
[tabby] ruins,” where along a dirt road he recovered a “number
of Spanish Olive Jar fragments, sherd tempered” (San Pedro)
and San Marcos wares, and two majolica sherds. Of the latter,
one was identied as Fig Springs Polychrome, which has a
production date of 1580-1650. The ruins are located in the
southern part of the Houston Plantation site (8DU90), which
is contiguous with the northern part of the Armellino site.
With the support of a grant from the Florida Division
of Historic Resources, UNF performed an intensive
archaeological survey of the southern third of Big Talbot
Island in the spring of 1998 (Ashley and Thunen 2000, 2008).
Approximately 450 acres were sampled on a tightly staggered
25 m grid to provide the State Park with information regarding
location, boundaries, and types of archaeological sites within
the park. UNF excavated 550 50-cm square shovel tests, which
resulted in the delineation of eight archaeological sites, seven
of which had been previously recorded but never bounded or
adequately sampled (see Figure 3).
The spatial boundaries of the Armellino site were
determined on the basis of 127 shovel tests, with the northern
and southern parts of the site tested mostly on a staggered
25-m grid; the site’s approximate center was sampled on
a rigid 25-m grid (see Figure 3). Ninety-eight (77 percent)
shovel tests yielded artifacts. A consistent, yet low density,
distribution of shell was encountered across the site,
punctuated by small, localized areas of dense shell midden.
The latter might represent household shell piles that were
leveled during nineteenth-century farming on the island. The
Armellino site is separated from 8DU90 to the north and
8DU628 to the south by small privately-owned outparcels
that were not sampled (permission not granted), meaning site
boundaries are articial. Moreover, the northeastern part of
8DU631 and site 8DU90 are distinguished from 8DU627 and
8DU80, respectively, by Houston Road; this dirt road runs
north-south between the private lots on the west and the large
wetland on the east.
Of the 234 recovered native potsherds, 49 (20.9 percent)
were Contact-era San Pedro and 27 (11.5 percent) were
early Mission period San Marcos. These wares were mostly
concentrated in the central part of the site, immediately north of
a private house lot, but a small cluster also was identied south
of the cemetery. Although San Pedro and San Marcos pottery
were widely distributed across the southern third of the island,
8DU631 appears to represent the core of the concentration
(Ashley and Thunen 2008:143). Thus, if the Contact period
village and later visita of Sarabay is located at the southern end
of the island, as we suspect, then the community and political
center of both may have been located within the boundaries of
8DU631. The proximity of this area to Mud River, a readily
navigable access route, lends credence to our contention that
the village center was located at the Armellino site.
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 55
Figure 4. Unit, block, and select shovel test locations, Armellino site (8DU631).
The Florida Anthropologist
56 2016 69 (1)
Unit 2 (4250N/1044E)
Unit 2 was placed 18 m east of Unit 1 within the same
private lot. The upper zone was black earth (10YR 2/1) about
30 cm thick. It was followed by a mottled gray (10YR 4/1)
and yellowish brown (10YR 5/6) sand matrix, exhibiting areas
of leaching from above. Five 10 cm levels were excavated
with the last two levels being saturated. Two liters of shell
were recovered from Levels 1 and 2, whereas 8 (10.9 g) pieces
of animal bone came from Levels 1-3. Nearly 90 percent of
the sherds larger than 2 cm derived from Levels 1-2; these
same levels accounted for 83.3 percent of the assemblage by
weight (see Table 2). The same pottery types recovered from
Unit 1 were identied in Unit 2. San Pedro was the dominant
type from Unit 2, composing a little less than 50 percent of
the assemblage by count and a little more than 50 percent by
weight. Only one San Marcos sherd was recovered.
Unit 3 (4284 N/1026E)
Unit 3 was placed 33 m north of Unit 1 on state park
property. This was the least productive of the three scattered
1-x-2 m units in terms of artifacts. About one quarter of the
pottery assemblage was San Pedro, whereas a little more than
half was St. Marys Cordmarked; the only other identied
types were St. Johns and sand tempered (see Table 2). Shell,
however, was more abundant in Unit 3, with Levels 2 and
3 producing a combined 13 liters. Twelve pieces of animal
bone (1.4 g) came from the same two levels. Five 10 cm
levels were excavated, revealing a series of sand zones that
included (from top to bottom): dark brown (10YR 3/3), very
dark brown (10YR 2/2), mottled dark brown (10YR 4/3) and
dark yellowish brown (10YR 4/4), and mottled grayish brown
(10YR 5/2) and dark yellowish brown (10YR 4/6). Level 4
yielded only four sherds while Level 5 was sterile. As with the
previous two units, Levels 4-5 were quite wet.
Block A-B General Levels
Block A-B, consisting of Block A (Units 4-12) excavated
in 1998 and Block B (Units 14-27) dug in 1999, was located
about 250 m south of Unit 1 (Figure 5). The same horizontal
bench mark was used throughout excavations so terminal
depths for each level were the same for each of the 23 units
composing the combined blocks. Four 10 cm levels were
excavated, and the subsoil interface was clearly visible at the
base of Level 4 (40 cm below surface). Although evidence of
plowing was observed in areas at the interface, the variable depth
and coverage of horse-drawn plowing with a small wooden
or metal plow failed to create a plow zone distinguishable
from the organic midden. All features and stains visible at the
interface were excavated in their entirety (discussed below).
Approximately 238 liters of shell were recovered from Block
A-B, with 74 percent coming from Levels 2 and 3. Shell was
scattered throughout the block, ranging from approximately 3
to 19.5 liters per unit, and was heaviest (greater than 15 liters)
in Units 22-24, 26, and 27 along the western and southern
Unit 1 percent Unit 2 percent Unit 3 percent
St. Johns 8
19.8g
5.6
2.8
3
13.8g
3.4
2.8
2
13.7g
3.0
4.2
Swift Creek 11
70.4g
7.7
9.9
1
2.3g
1.1
0.5 - -
St. Marys 79
380.0g
55.2
53.3
16
80.1g
18.0
16.2
35
165.2
52.2
50.5
San Pedro 17
102.9g
11.9
14.4
42
256.8g
47.2
51.8
17
93.9
25.4
28.7
San Marcos 8
63.5g
5.6
8.9
1
5.7g
1.1
1.2 -
Sand tempered 20
76.0
14.0
10.7
26
136.7
29.2
27.6
13
54.5g
19.4
16.7
Total 143
712.6g
100
100
89
495.4g
100
100.1
67
327.3g
100
100.1
Sherd < 2cm 212
228.9g
122
144.5g
180
203.4g
Shell Volume (liter) Trace ~2 ~13
Animal bone 15
7.2g
8
10.9g
12
1.4g
Table 2. Units 1-3, summary of pottery, animal bone, and shell (count/weight in
grams).
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 57
Figure 5. Block A-B and Shell Midden A units.
edges of Block B. Oyster was the most common species with
much smaller amounts of broken quahog clam, Atlantic ribbed
mussel, stout tagelus, and whelk.
San Pedro and San Marcos pottery were recovered together
in the upper 40 cm. These two wares accounted for 85 percent
of the block ceramics in terms of both count and weight (Table
3). San Pedro was the dominant ware in each excavation level.
With respect to the total Block A-B assemblage, San Pedro
outnumbered San Marcos by a ratio of almost 2-to-1. However,
relative to San Pedro, San Marcos was much more common in
Levels 1 and 2. Other pottery types relevant to the Contact and
early Mission periods were Irene-like sherds and colonowares.
The former had heavy, coarse to very coarse grit temper and
carinated incised shoulders, whereas block specimens of the
later were native-made, brimmed red-lmed plates or bowls.
Figures 6 and 7 show a sample of recovered pottery types,
while Table 4 provides a breakdown of the major pottery types
from the blocks by paste characteristics.
Another ceramic artifact recovered from Unit 11 (Level
3) was a fragment of an incised tobacco pipe. Specically, a
section of what Blanton (2016:95-96) refers to as a stemless
Citico-style pipe, which is strongly specic to the central
Georgia coast. The Armellino specimen depicts an ear element
and as Blanton (personal communication, 2012) states this
pipe style “varies slightly but also very consistently presents
a particular package of elements. In effect, the head/face
emulates the central serpent head typical of Citico-style
gorgets. The ‘broken line’ element below your ear [Figure 8] is
most likely the bared teeth of the gure. This style is securely
dated to very late prehistoric-early contact eras.”
The general levels of Block A-B yielded a small collection
of lithic material that included 10 pieces of chert (akes and
shatter), three rounded, pink or white quartz pebbles, two
sandstone fragments, a biface tip, and a broken Pinellas point.
Because of plow disturbance, it is unclear to which time period
these date, but it is likely that some are associated with the
site’s Contact component.
Unequivocal evidence of post-contact interaction with
Europeans exists in the form of 18 olive jar sherds, 3 hand-
wrought nail fragments, a cut fragment of lead shot, an engraved
brass nial, and a few pieces of lead and miscellaneous iron
recovered from the four excavation levels (Figure 9). Nails
and other miscellaneous iron were most prevalent in the
northern units of Block B, while the 18 olive jar fragments
were spread along the northern and eastern units of Block
B. The brass nial was likely attached to end of a piece of
leather, such as a belt or scabbard. According to John Powell,
who is responsible for identifying many of the military objects
recovered by St. Augustine City Archaeologist Carl Halbirt,
the item is a scabbard tip
The Florida Anthropologist
58 2016 69 (1)
L-1 L-2 L-3 L-4 Total Count
Percent
Weight
percent
Orange - - - 1
2.8g
1
2.8g .04 .02
St. Johns 4
12.1g
9
36.6g
9
28.0g
4
7.6g
26
84.3g 1.2 0.6
St. Marys 2
5.1g
13
52.4g
12
46.2g
24
112.7g
51
216.4g 2.3 1.6
San Pedro 185
955.2g
382
2185.9g
464
2762.0g
242
1699.9g
1273
7603.0g 56.1 55.7
San Marcos 127
643.7g
257
1580.6g
243
1616.5g
38
281.0g
665
4121.8g 29.3 30.2
Irene 2
13.7g
18
89.5g
11
104.5g
1
5.2g
32
212.9g 1.4 1.6
Colonoware 1
7.4g
9
58.6g
15
90.2g
2
12.9g
27
169.1g 1.2 1.2
Sand tempered 25
107.2g
28
138.4g
71
326.4g
40
215.0g
164
787.0g 7.2 5.8
Grit Grog tempered 3
15.7g
-2
11.8g
5
36.1g
10
63.6g 0.4 0.5
Grit tempered - - - 1
12.4g
1
12.4g .04 .09
Olive Jar 6
96.5g
10
257.0g
1
5.8g
1
25.8g
18
385.1g 0.8 2.8
Total 355
1856.6g
726
4399.0g
828
4991.4g
359
2411.4g
2268
13658.4g 99.98 100.11
Sherd < 2 cm 519
711.4g
1612
2120.0g
1668
2172.4g
784
950.8g
4583
5954.6g
Shell (liters) 28 82 94.5 34 238.5
Bone 75
31.8g
395
136.6g
843
312.4g
647
273.4g
1960
754.2g
Iron Oxide 43
66.6g
208
280.7g
296
471.0g
163
239.2g
710
1057.5g
Table 3. Block A-B, Levels 1-4 (count/weight in grams).
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 59
Figure 6. San Pedro sherds (top: cob-marked and check stamped; bottom: fabric and cob-marked, both partially
obliterated.
…likely from a Spanish sword or main gauche
(parrying dagger). Unlike hilt designs, scabbard tips
remained the same in style for very long periods of time
and are difcult to precisely date because uniformity in
military edged weapons did not appear until after this
was made. The punch designs at the neck of the piece
are obviously hand done and very Spanish in detail, and I
would place the date of this to the 17th century to very early
18th century. I wouldn’t call it truly military because both
soldiers and civilians carried these weapons and there was,
as I mentioned, little uniformity except in overall style and
function (John Powell, personal communication, 2011).
Of special note was the recovery of inordinate quantities
of iron oxide (hematite) nodules scattered across the two
blocks. In fact, all units contained sandy or low-sand, rounded
or subangular nodules of iron oxide. In all, 710 fragments
weighing 1057.5 g were recovered from the blocks (see Table
3). Analysis of non-feature unit data revealed no clustering of
iron oxide, which might have suggested specic activity or
processing locations within the excavation blocks. Evidence
reveals a greatest volume in Level 3 (44 percent of total
weight) followed by Level 2 (27.1 percent), Level 4 (21.8
percent), and Level 1 (6.2 percent). Iron-rich oxide is not local
to the area and was likely brought in to be ground for use as red
pigment often observed on the interior surfaces of colonoware.
Interestingly, very little red lmed pottery or fragments of
colonowares occur in the block assemblage.
Vertebrate faunal bone was lightly dispersed throughout
the four general levels in all 23 units, and its weight ranged
from 11.6 to 90.0 g per unit. Save for Unit 16 (53.6 g) and Unit
10 (90.0 g), the remaining units each yielded less than 50 g of
bone. In Units 10 and 16, deer and large mammal bones from
Level 4 accounted for the higher weight of recovered bone.
The low numbers and weights of bone likely reect a variety
of factors such as discard behavior (outside the primary living
area), matrices lacking shell as a preserving agent, and a
native focus on small sh as a dietary staple. The majority
of subsistence bone is concentrated in Levels 3 (n=848, 42
percent; 396.0 g or 50 percent) and 4 (607, 30 percent; 269.2 g
or 34 percent). The lowest bone weight was recovered in Unit
9 (n=64, 11.6 g or 1.6 percent of the unit weight totals), which
The Florida Anthropologist
60 2016 69 (1)
Figure 7. San Marcos red lmed colonoware (top row), San Marcos (lower left), and Irene (lower right).
is adjacent to Structure 1 in the southwestern section of Block
A. The average bone weight per level was 34.2 g (median was
29.9 g).
In terms of specic remains recovered, the majority of
bone in the general levels came from sh that easily could
have been captured in nearby Mud River: catsh, mullet,
drums and ounder. Medium-sized bird long bones also were
present along with a variety of turtles and snake. Rabbits and
other medium-sized mammal bones were all highly fractured.
Deer teeth, cranial fragments, and the distal and proximal ends
of long bones were found in the general levels. Immature and
fused mammal bones were observed. Few of these bones were
burnt. One pig bone was recovered from Unit 11.
Twelve (52.2 percent) units yielded modied bone. The
weights are low, reecting the splintered and fractured nature
of bone reduction for tools and tool failure. Modied bone was
recovered in Level 2 (n=22, 28 percent; 11.4 g or 20 percent),
Level 3 (n=36, 46 percent; 28.5 g or 51 percent), and Level
4 (n=18, 23 percent; 14.5 g or 26 percent). By weight, Unit
9 (25 percent), Unit 19 (16 percent), and 20 (15 percent)
contained the greatest concentration of modied bone. Very
little modied bone is associated with the immediate area of
the proposed structure.
Of the identiable modied bone, at least three animal
classes provided raw material for tools. Four polished bird
long bones (0.4 g) from medium-sized species were recovered
(Units 19 and 20). All were polished, and one exhibited groove
and snap marks at the bone ends. One polished sh spine bore a
damaged tip (Unit 20). Eight turtle carapace fragments showed
polish on the shell interior; two of these also bore polished
exteriors (Units 9, 19, 20, 23, and Area 14). Fourteen long
bone slivers could clearly be identied as deer. Eight of these
displayed heavily polished interior surfaces or rare exterior
polish. Four exhibited a familiar repeated pattern of battered
base, edge lash scar and areal surface abrasion, and polish at
battered pointed or beveled tips. The 52 remaining modied
bone fragments reveal combinations of traits: polished interior
and/or exterior surfaces, smooth-edge beamers, knapped
pointed or c-shaped tips, and one with two drilled holes. Two
fragments of incised pins were present.
In terms of shell modied whelk, clam, or giant Atlantic
cockle were recovered in 18 (78 percent) of the Block A-B
units. Modied shell includes whelk fragments from reduction
with or without obvious edge wear. Fourteen columella tools
were recovered in a wide range of sizes, implicating the need
for ne punch to blunt battering tools. Eight whelk specimens
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 61
San Pedro
Fine to coarse sand tempered with crushed grog (prered clay)
Frequency of grog is highly variable and grog particles maybe
angular (larger pieces) or rounded
Frequent to abundant very-ne to ne quartz inclusions
Rare to occasional coarse quarts inclusions
St. Marys
Fine Texture dominated by very ne-, ne, and medium-sized
quartz Inclusions
Occasional coarse-sized particles; Rare grit-sized inclusions
Occasional to frequent ne to medium-sized mica platelets
San Marcos Range of quartz inclusions that varied in frequency
Irene-like
Common to abundant, angular or subangular, coarse to very
coarse quartz grit temper
Reduced ring with hard tooled to burnished nishes
Combination of vague complicated stamping below carination
with parallel-line incising above
Simple hemispherical or slightly excurvate rim forms
Sand tempered decorated (including uid)
Medium to coarse texture
Occasional very-ne to ne quartz
Frequent to abundant medium- to very coarse-quartz inclusions
Occasional grit-sized quartz inclusions
Probably low clay particle base as nger pressure results in edge
attrition and sherd friability
Sand tempered plain
Fine to medium texture
Occasional to common very ne- to ne-quartz inclusions
Frequent to common medium- to very coarse- quartz inclusions
Finger pressure does not bring about edge attrition
Colonoware Sand, grog or grit inclusions may be observed in a wide range of
frequency
Table 4. Paste descriptions, 8DU631.
bore severely damaged tips. Separated outer whorl walls
and whorl lips showed erosion and smoothing. One small
unidentied gastropod columella was recovered in Unit 5. It
exhibited one battered tip. One giant Atlantic cockle fragment
(Unit 8) had been reduced to a small rectangular tool. Though
recovered in low numbers, this tool type made of the same
species is found at many regional sites. Eleven clams with
narrow spokeshave-like lip indentions were recovered in Units
15, 16, and 17 and Area 14.
Charcoal was common throughout the excavation
blocks, but was not subject to detailed analysis. Hickory
hulls, unidentied seeds and beans were noted, as were 11
cob fragments from Unit 16 in the north central part of the
Block A-B. Corn was also recovered from Unit 23. Because
of its friable nature, charred maize was more prevalent (and
protected from plowing) in features, as discussed below.
Block A-B Features
A total of 52 cultural features was identied and
excavated in Block A-B (Figure 10). Each was determined
to be cultural based on its conguration in plan and prole,
which exhibited distinct and well dened edges in comparison
to the surrounding soil matrix. Generally, feature shapes and
plans were regular and rounded. Because of the large number
of features, we only describe the broad categories of features
encountered in the block. Appendix A provides metric data
and descriptions of each feature. Figure 11 depicts selected
feature proles along block walls.
Pits (n=19) are holes of varying size, shape, and depth
dug by site occupants. Most pits were concentrated in an area
slightly north and east of Structure 1. These were typically
oval to round in plan with inward sloping sides and a round
bottom, giving them a bowl shape in cross section. Excluding
Feature 26, which was only partly exposed, excavated pits
ranged in length from 22 to 55 cm, in width from 18 to 42
cm, and in depth from 9 to 31 cm.2 The average pit size was
34.1 by 27.9 by 20.7 cm deep. Some were rather small and
shallow and may have been dug as small roasting pits. Many
were utilized as receptacles for trash, although this activity is
thought to have occurred after the pit had ceased to serve its
original function (e.g., storage, processing, roasting). Thus,
judging from their size, abundance, and presumed short use
life, these expediently constructed pits were associated with a
variety of storage and/or processing tasks.
Postmolds/holes (n=26) tended to be round in plan,
although when originally encountered some were somewhat
The Florida Anthropologist
62 2016 69 (1)
Figure 9. Brass scabbard tip (top left), Mexican Red Film (top right), olive jar (bottom row).
Figure 8. Native pipe bowl, various views.
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 63
Figure 10. Plan of Block A-B showing features at the plow zone-subsoil interface.
The Florida Anthropologist
64 2016 69 (1)
Figure 11. Select wall proles, Block A (note features).
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 65
distorted in plan due to plowing. With depth, however, a
circular plan outline became evident. In cross section, most
postmolds tended to have parallel walls (sides) that ended
with a slightly rounded bottom; a few had at bottoms or
pointed bases. The majority of posts appear to have ranged
from approximately 15-25 cm in diameter, and from 4 to 62
cm in depth (the average depth is 34.6 cm). Artifacts from
these features are largely assumed to have been inadvertent
inclusions.
Other features included two charcoal concentrations
(Features 41 and 51), one of which contained charred corn
cobs. In addition, a large and relatively shallow area of gray-
brown mottling (Feature 20) that appears to represent the
entrance to Structure 1 and a diffuse area of burning (Feature
36) in the northwestern corner of the Block B.
The most conspicuous feature was an arc-shaped stain in
the extreme southwestern corner of Block A, which appears to
be an undisturbed basal section of a wall-trench, designated
Structure 1 (see Figure 10). The trench averages about 25 cm
in width and ranges from 15-20 cm in depth. Based on the
curvature of the exposed section of the wall trench, the feature
suggests the outline of a circular building with an estimated
diameter of approximately 6.7 m (22 ft); however, an oval
shape is also possible. Using this projected size, it appears
that a little less than one-fourth of the structure’s outer wall
was contained within Block A and excavated. Unfortunately,
additional excavations to reveal more of Structure 1 were not
possible due to the presence of large trees immediately outside
the southeastern corner of Block A.
In terms of construction, the structure seems to have
consisted of small-diameter posts placed in a wall trench. The
absence of daub suggests that the outer wall was likely built of
wattle and thatch or palmetto. Because the structure was not
burned, none of the organic materials used in its manufacture
has preserved. A visible gap in the wall trench, manifested as
a thin, amorphous area of mottled soil and rare broken and
crushed shell (Feature 20), points to the presence of a door or
entryway. This opening seems to rule out the possibility that
the feature served as a screen. At this point, we cannot dismiss
the prospect that the building combined trench and single
post construction. Because the feature was not excavated
in its entirety, we are limited in what we can say about the
conguration of this hypothesized structure. No evidence
was uncovered to suggest that the building had a clay oor or
any kind of prepared surface. In addition, no indication of an
internal hearth or re pit was found, although such a feature
may have been more centrally located within the building
beyond the area of excavation. The same can be said of the
absence of internal roof support posts for a rigid structure.
Features within the exposed interior section of Structure 1
include three pits and one posthole. Collectively, these yielded
corn, olive jar, and high quantities of iron oxide, including
one large block in Feature 23. The latter, which weighed 250
g, was recovered from what appears to have been a posthole,
although it may have been a straight-sided pit for storing or
caching hematite. Also encountered just inside the eastern
wall of the proposed structure (and likely abandoned on the
structure oor) was a nearly complete San Marcos vessel
(Feature 6) lying on its side (Figure 12). The underside of
the vessel was intact, whereas the upper side was scattered
into sherds over an area of a few meters, likely the result of
nineteenth-century, horse-drawn plowing. Inside the vessel
was a complete Giant Atlantic cockle shell with a small square
perforation. Repeated etching associated with the removal of
the square extends beyond the actual hole. Most of the sherds
representing the vessel came from Levels 3 and 4. The San
Marcos vessel and olive jar sherd tentatively date the structure
to the early Mission period, ca. 1600-1610.
North of the structure was an array of postmolds and pits
(see Figure 10). In terms of the former, a roughly SW-NE
alignment exists across the block from the eastern-central part
of Block A to the northeastern corner of Block B. A roughly
perpendicular distribution extends from the approximate
center or this alignment to the northwestern corner of Block
B. Other postmolds are peppered throughout the northwestern
quadrant of the larger block with fewer located in the southern
third of the Block A-B. It is difcult to determine if these are
associated with a structure(s) but the possibility does exist.
The most conspicuous distribution of features is an arc-shaped
dispersal of pits, immediately northeast of Structure 1. Finally,
within, the northwestern corner of Block B is an amorphous
area of dark soil and charcoal, apparently the byproduct of
some form of re. Not enough of the burned area was exposed
to interpret what had occurred. As Figure 10 demonstrates,
few features were revealed in the western and most eastern
parts of Block B.
San Pedro was the primary pottery type recovered from
the features (Table 5); it occurred in 33 (63.5 percent) of the 52
features in Block A-B. It was most prevalent in the northwestern
portion of the block: Features 8 (223.2 g), 12 (43.3 g), 15
(151.3 g) 26 (93.1 g), 27 (99.3 g), 31 (93.7 g), 36 (86.8 g),
37 (65.8 g), 38, (96.4 g), and 49 (90.4g). Small amounts of
St. Marys were recovered from 12 features. Ten of these also
yielded higher quantities of San Pedro pottery. Sand tempered
plain, along with a few cordmarked, fabric marked and cob
marked were included in feature ll. It is likely that most of
the St. Marys and sand tempered wares were incorporated into
San Pedro features. A single colonoware sherd was recovered
from Features 36, 39, 43, 46, and 49; all but Feature 43 are
located in the northernmost units. Interestingly, no San Marcos
pottery was recovered from any of the features.
Feature 19, however, yielded Spanish olive jar. Feature
36 contained a broken celt fragment used as a hammer or
pounding tool. A single piece of chert was recovered in Features
19, 20, 22, and 36. Three of these features lay in the extreme
southwestern units. Chert artifacts include a thermally altered
ake, two pieces of shatter, and a bifacially retouched stem
fragment. A single smoothed piece of sandstone that may have
been used as an abrader or hone was recovered from Feature
13, a pit near the structure. A tiny fragment of fossilized turtle
shell (Feature 20) bore a highly polished surface.
Iron oxide was recovered in 30 features (60 percent).
Represented by small rounded or angular lumps, it was widely
distributed throughout the entire block. Higher concentrations
The Florida Anthropologist
66 2016 69 (1)
Figure 12. Feature 6, San Marcos vessel containing giant Atlantic cockle shell.
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 67
bone was recovered from 19 features (37 percent).
A single snake vertebra (Feature 12) came from
a non-poisonous species. The majority of turtle
bone consisted of one or two carapace fragments
that had come from small to medium-sized
aquatic turtles (no gopher tortoise). Bird bones
were recovered in 16 (25 percent) features and
evenly represented in both pit and post features.
Bird skeletal elements represented a complete
range of body sizes from small, medium (gull-
sized) to large (pelican- and vulture-sized).
Other discarded elements include low-meat
portions of the body: vertebrae, tibiotarsus, distal
humerus, ulna, and carpometacarpus. Rather than
subsistence, perhaps feathers were the focus of
their capture.3
Modied bone was recovered in 17 features (32 percent),
which were scattered across the block units. Seven (41
percent) of the modied bones were recovered in pit features
while the remaining ten had been discarded in posthole
features. Of the 37 bones, 36 exhibited informal aboriginal
use-wear modications which are listed in Table 7. The single
formal object—a broken shing hook with a line-attachment
groove—was recovered in Feature 37. Utilized bone included
at least three faunal classes: bird (n=3, 8 percent), bony shes
(n=3, 8 percent), and mammals (n=24, 65 percent). Another
seven pieces (22 percent) could not be securely identied to
class, although medium to large mammal seems their likely
classication.
While bird bone was recovered in 16 features (30 percent)
the three polished bones were deposited in Features 9 (pit
feature) and posthole Features 31 and 50. Three polished sh
pectoral spines, including 1 catsh, were found in posthole
Features 12 and 52 (n=2). Spine tips were battered and
missing. Twenty-four mammal bones (deer, and unidentied
medium to large) exhibited single or multiple evidences of
tool production, use, or structural failure. Five mammal long
bones had been “knapped” to form a pointed tip and exhibited
either polished compact bone or lashing scars (perpendicular
to bone length) which were bordered by akey-textured
abraded surfaces. Pointed-tip long bone punches often bore
smoothed interior edges along the tapering sides and tip. Less
often these piercing tools appeared with jagged or battered
tips suggesting failure at a narrow pressure point. A small
collection of ve polished large- and medium-sized mammal
(by weight) were found in Feature 3 (188.1 g), Feature 8
(22.2 g), Feature 19 (12.9 g), and Feature 22 (17.1 g). Three
of these lie immediately north of Feature 23, which yielded
a 10 cm3 block of iron oxide weighing well over a kilogram.
The four features are closely associated with Structure 1. Iron
oxide can be used in the production of red lm associated with
native colonoware vessels, although only three of the very
small ve colonoware sherds from the block were red-lmed.
This reects a signicantly lower presence of red lming or
colonoware compared with that recovered from the nearby
missions of Santa Cruz and Santa Catalina de Amelia (Ashley
et al. 2013b; Rolland and Ashley 2000).
The majority of the Block A-B features contained organic-
preserving shell and/or charcoal. Bones were recovered in 41
(77 percent) of the block features. All bone greater than ¼” was
analyzed, whereas samples of sub-¼” bone were scanned for
the presence of identiable species. The number of identied
specimens (1/4”) was 3509, which produced a total weight of
only 178.8 g (Table 6).
Fish was the most common class. Fish body sizes ranged
from ngerlings to adult, although large adults were less often
observed. The assemblage contained bone from 18 species of
sh typical to the adjacent tidal waters. Freshwater rays also
were present. Mammal bone was deposited in 21 (40 percent)
features. Juvenile and adults were represented, as well as
examples of small-, medium-, and large-sized species. Reptile
Count Percent Wt (g) Percent
St. Marys 13 7.6 69.4 4.4
San Pedro 128 74.4 1412.2 89.0
Colonoware 5 2.9 8.7 0.5
Sand tempered Decorated 13 7.6 44.5 2.8
Sand tempered Plain 13 7.6 52.3 3.3
Total 172 100.1 1587.1 100.0
Sherd < 2 cm 363 369.6
Grand Total 535 1952.2
Table 5. Pottery totals from features.
Class Count Percent Wt (g) Percent
Fish 2773 79 64.0 36
Bird 34 1 12.5 7
Reptile 130 4 17.9 10
Amphibian 1 0 .01 0
Mammal 62 2 57.2 32
Crab 35 1 1.7 1
Unidentied 474 13 25.5 14
Total 3509 100 178.8 100
Table 6. Unmodied bone from features by faunal class.
Type of Use-wear Count Percent Wt (g)
Exterior polishing 13 35 25.1
Cut marks 1 3 0.6
Knapped to tip or punch 5 5 8.5
Lash scars, battered exterior 3 8 18.9
Fish hook 1 3 0.9
Table 7. Informal and formal modied bone from features
(more than one condition may apply).
The Florida Anthropologist
68 2016 69 (1)
bone was recovered in Feature 15 (post feature). Cut marks
were rarely observed, and utilized bone in this assemblage was
rarely thermally altered.
A variety of invertebrates were recovered from the features.
Crab claws (mostly burnt) were identied in 14 features (27
percent). As is typical for barrier island sites in northeastern
Florida, oyster was the most commonly exploited shellsh
species in the features followed by quahog clam, stout tagelus,
and Atlantic ribbed mussels. The numbers of marsh periwinkle
and mud nassa (probably collected alongside Atlantic ribbed
mussel) were lower but constant across the excavation
blocks. Ponderous and blood ark species were occasionally
encountered but none exhibited a perforated umbo associated
with shing-net technology. Coquina was a minority species
that was mixed in the ll of six features, largely concentrated
in the extreme southwestern units. Coquina live in the ocean
tidal zone so an obvious effort had been made to travel to and
transport this species back to the site.
Modied clam shells (n=19, 34.8 g) were recovered in
ten features (20 percent). There was no areal concentration
of utilized shell that again might suggest a shared activity
area. The majority were reduced shells (1/4 to 1/3 of valve)
suggesting a size better suited for manipulating cordage or
processing bone. Both marginal wear and C- or V-shaped
spokeshave damage was recorded. Polish around the umbo
exterior was found in lower numbers.
Only ve whelk (whole n=3, 794.2 g or debitage n=2,
3.0 g) were recovered from three features (15, 37, and 47).
One knobbed and one Keiner’s whelk had been hafted and
appeared to have been discarded in still usable condition. The
third whole shell was a smaller knobbed whelk hand tool with
posterior and node smoothing. It exhibited lip reduction back
behind the rst nodule and wear-perforation into the second
tier of the spire above the aperture. A single large nodule and
wall fragment also were recovered.
Giant Atlantic cockle (n=25, 41.31 g) is another marine
species utilized as small hand tools. As shell tool research
continues in this region, we have noted the reoccurrence
of reduced, similarly-sized and shaped fragments of giant
Atlantic cockle. This species was recovered in 14 features (27
percent). Such fragments were few in number within a total
shell assemblage and often found deposited singly. We have
begun collecting all cockle fragments and are beginning to
see a repetition of fragment size, distal edge wear, and interior
areal polish. In Features 6 and 13 we recovered two whole
Atlantic cockle shells. As mentioned previously, the Feature 6
shell was found within a San Marcos vessel, while the second
cockle was recovered from a large pit designated Feature 13.
One shark eye (a marine species) columella was recovered
in Feature 10. A second modied shark eye, showing lip and
whorl reduction/abrasion, was found in Feature 45. Tiny
marine rice olive shells (n=94, 12.0 g) were recovered in nine
features (16 percent). The highest number were found in post
Feature 34 (n=79, 5.3 g). Some individual shells had had their
apex removed, perhaps for stringing. As their nutritional value
seems negligible, their select presence at this site suggests
a new commodity specically collected. Lastly, a single
unmodied disk shell was also recovered from Feature 34.
Although a formal ethnobotanical analysis was not
undertaken a few general comments can be forwarded. Whole
seeds were recovered in large and small screen fractions or
otation samples from 14 features (2 percent). Fragmented
monocots and dicots were noted. Maypop seeds, which were
rare, and hickory hulls, were the only native species we
could identify. Corn was recovered in 33 features in Blocks
A-B. Because of the extreme friability of cobs and kernels,
the counts and weights of these remains are only marginally
useful in interpreting the importance of corn to the Sarabay
population whereas horizontal distribution conrms its
importance. There is a somewhat greater concentration of corn
in the extreme southwestern units: Feature 4, 0.7 g; Feature
15, 0.8 g; Feature 22, 0.6; and Feature 23, 0.6 g. The greatest
cache was recovered in the charcoal concentration of Feature
41 (2.3 g). A cob from Feature 23 was AMS dated to A.D.
1450-1610.
Shell Midden A
Moving away from Block A-B, Units 13, 28, and 29 were
placed in Shell Midden A, approximately 20 meters to the
northeast (see Figure 5). Units 28 (1-x-2 m) and 29 (1-x-1 m)
were combined to form a 1-x-3 m unit, while Unit 13 was
a one-meter square placed about one meter to the southeast.
Unit 13 was actually an expansion of Shovel Test (ST) 545,
a 50 cm square unit initially dug to sample Shell Midden A.
Units 13, 28, and 29 are considered herein as one analytical
unit. ST 545 produced 14 sherds, 9 (66.8 g) San Pedro and 5
San Marcos 5 (27.9 g), which are excluded from the following
discussion.
Shovel testing of the southern third of Big Talbot Island
had revealed a rather spotty distribution of San Pedro period
refuse. Although few distinct shell heaps were observed during
transect testing, it is likely that once mounded middens had
been leveled and scattered as a result of a century of farming
during the late eighteenth and nineteenth centuries. Although
it was difcult to discern a distinct rise in Shell Midden A due
to its forested setting, probing with a metal rod indicated that
it measured about 13-x-11.5 m.
Excavation of six 10-cm levels revealed a moderately
dense midden with approximately 216 liters of shell removed
from the three units. Ninety-one percent of the shell was
recovered from Levels 1-3. A high percentage of heavily
fragmented oyster with fewer whole shells (minority species
included tagelus, mussel, and quahog clam). The ¼” faunal
sample included 229 bone fragments (66.5 g) consisting
mostly of sh, turtle, and deer. San Pedro was the dominant
ware comprising 82.9 percent of the recovered ceramics
and outnumbering San Marcos 107 to 8. By weight San
Pedro accounted for 85.8 percent of the ceramic assemblage
compared to the 9.2 percent of San Marcos (Table 8).
A fragment of olive jar was taken from Unit 29 and a
piece of Mexican Red Filmed occurred in Unit 28. Iron oxide
fragments (138.9 g) were recovered from all three units,
being most prevalent in Levels 2 and 3 of Unit 29. A small
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 69
retreat from the Georgia coast during the seventeenth century
(John Worth, personal communications, 2012). A common
characteristic of all suspected contact villages and outlying
visitas in northeastern Florida is a low-frequency occurrence
of European goods. The most commonly received item from
the Spanish was cloth (or clothing), which typically does not
preserve in archaeological contexts. Early on other items
may have accompanied the deceased to the grave. The most
regularly recovered European artifact on historic Mocama
sites in the region is pottery, mostly olive jar. More than 30
fragments of this Iberian ware have been recovered from the
southern end of Big Talbot Island since the 1960s.
Second, the primary method of refuse disposal among the
Mocama was in the form of individual shell heaps peppered
over broad areas (refuse was also deposited in pits after
they had outlived their usefulness). This combined with the
widespread distribution of San Pedro pottery across suspected
village sites suggests a more diffuse community layout than
the tightly clustered, walled-village depicted in the Theodor de
Bry engraving of a Timucuan village. It is becoming clear that
this out-of-place and overplayed image has its roots far from
the Florida coast (Feest 1988; Hann 1996:88; Milanich 2005;
Sturtevant 1992).
Third, corn is a very late addition to a predominately
coastal shellsh collecting-shing-hunting domestic
economy. All available radiometric dates (at the two sigma
fragment of rusted iron from Unit 28 Level 2, a piece of chert
shatter from Unit 13 Level 4, and two fragments of modied
whelk from Unit 13 Levels 2 and 4 round out the non-ceramic
artifacts.
Three features were encountered at 48 cm below surface
in Unit 13 at the shell midden-subsoil interface. Feature 28
was a small (18 x 22 cm) basin-shaped feature (12 cm deep)
containing burned corn cobs. Cob fragments were also found
in the base of Level 4 directly above the feature. Feature 29
was a circular (24 cm in diameter), bowl shaped (20 cm deep)
pit that contained charcoal ecks and chunks and crushed
shell (7 liters) in a dark grayish brown sand matrix. Feature
30 appears to represent a dip or low spot in the overlying shell
midden. Finally, a possible postmold was identied in Unit 29.
Discussion and Conclusion
What have we learned from our work at the Armellino
site? First, there is no doubt that the Armellino site dates to
the historic period. We believe documentary, cartographic,
and archaeological evidence supports the contention that the
south end of Big Talbot Island is the location of the Contact
period village/visita of Sarabay. Based on all available
documents, no native community is known for Big Talbot
Island except Sarabay. Neither is any relocated Guale mission
community known to have settled on the island during their
Unit 13 % Unit 28 % Unit 29 % Total %
St. Johns - - 4
18.9g
5.8
3.2
2
5.6g
6.1
2.8
6
24.5g
4.7
2.5
St. Marys - - - - 3
7.3g
9.1
3.7
3
7.3g
2.3
0.7
San Pedro 22
156.8g
81.5
71.6
61
539.1g
88.4
92.0g
24
163.8g
72.7
83.1
107
859.7g
82.9
85.8
San Marcos 4
59.2g
14.8
27
3
24.6g
4.4
4.2
1
8.6g
3
4.4
8
92.4g
6.2
9.2
Sand tempered 1
2.9g
3.7
1.3
- - 2
7.2g
6.1
3.7
3
10.1g
2.3
1
Mexican Red Film - - 1
3.4g
1.5
0.6
- - 1
3.4g
0.8
0.3
Spanish Olive Jar - - - - 1
4.6g
3
2.3
1
4.6g
0.8
0.5
Total 27
218.9g
100
99.9
69
586.0g
100.1
100
33
197.1g
100
100
129
1002.0g
100
100
Sherd < 2cm 23
28.0g
117
159.5g
58
58.5g
198
246.0g
Shell Volume (liter) 39 ~94 ~83 ~216
Animal bone 86
35.4g
104
21.9g
39
9.2g
229
66.5g
Table 8. Unit 13, 28-29, summary of pottery, animal bone, and shell (count/weight in grams).
The Florida Anthropologist
70 2016 69 (1)
level) indicate that maize was incorporated into the Mocama
subsistence economy no earlier than A.D. 1450 (Ashley
2009:131). Its rst appearance in the region is coincident with
the emergence of San Pedro pottery, which includes a cob-
marked type. Interestingly, similar late dates on corn from
Irene (Guale) sites along the north Georgia coast suggest that
maize cultivation is much later along the Atlantic seaboard
than previously suspected (Thomas 2008:1099-1100).
Fourth, the presence of lesser amounts of San Marcos
pottery on sites with major San Pedro components points to
an in situ ceramic change and to continuity in the location
of Contact period villages and early Mission period visitas.
Because the last mention of the visita of Sarabay in Spanish
documents is 1602, the possibility exists that San Marcos
was being produced by the local Mocama as early as the rst
decade of the seventeenth century. Block A-B yielded a much
higher percentage of San Marcos pottery than Shell Midden
A, suggesting that the shell midden may have been deposited
early in the transition from San Pedro to San Marcos (Table
9). The same may be true for the lower levels of Block A-B,
where San Marcos was absent from features.
Perhaps the most intriguing feature uncovered at the
Armellino site is one that leaves us with the most questions
due to the incompleteness of its exposure during eldwork.
Structure 1 appears to represent a circular (or oval) wall trench
building estimated at 6.7 m in diameter. Slightly less than one-
fourth of the wall trench was excavated and an even smaller
section of the interior was investigated. To date, in contrast to
Spanish mission structures, there is a dearth of archaeological
information on domestic structures from late prehistoric and
Contact period sites in northeastern Florida. While features
such as shell middens, hearths, refuse pits, burials, and possible
postholes have been excavated at several sites, complete or
even condently discernible posthole patterns remain elusive.
Wall-trenches are known for late prehistoric and contact sites
(Irene) along the north Georgia coast (see Keene and Garrison
2013 for a review of Irene phase architecture). These all are
square or rectangular, although a few might have had rounded
corners. Several of these were heavily daubed with clay.
Circular wall trench structures are not well reported
for coastal Florida or Georgia. However, Milanich (1973)
excavated an oval wall-trench house that measured 32 by 14
feet (9.8 X 6.7 m); a size comparable to Structure 1. Located
on near the northern end of Cumberland Island, Georgia, this
structure dated to the early Deptford period. This “house”
contained a door, an interior slot trench partition, two possible
support posts, and a large-centrally located re pit (Milanich
1973:107-112). A whelk shell pick from the re pit generated
an uncorrected radiocarbon date of A.D. 55±95. Portions of a
burned wall trench structure were exposed beneath the Queen
Mound, a Woodland period Deptford burial mound along the
lower St. Johns River (Lafond and Ashley 1995:38-39). Details
are lacking, but the structure appears to have been square or
rectangular. Both of these structures predate the Armellino site
by more than one millennium.
The presence of a wall trench implies a rigid structure
with solid walls. Buildings like this are known for the
Mississippian Southeast, where most are viewed as sweat
lodges or storage facilities (Gibson 1998:870). However,
these buildings are typically much smaller than the projected
size of Structure 1. The Armellino building might represent a
large domestic structure or a specialized facility for storage or
meeting. Since Structure 1 was not completely excavated, we
lack solid information on associated interior features and its
ultimate conguration and function.
In sum, excavation of the Armellino site on Big Talbot
Island generated a wealth of data on the Contact and early
Mission period Mocama Timucua of coastal northeastern
Florida. The largely native occupation of the Armellino site
provides an introduction to our understanding of the spread
and adaptation or indifference to European ideas and goods
in outlying villages turned visitas. One stark example is the
concentration of iron oxide unlike that recovered in any other
Mission-period sites in the immediate region. Does this suggest
that Sarabay served as an independent economic center for the
production and distribution of this particular material desired
for Spanish dinner wares? And, although the quantity of iron
oxide was high, native potters chose not to adopt the step
of lming their own vessels nor were Spanish vessel forms
recovered in signicant numbers. It is also likely that the red
coloring was used on other items and perhaps buildings or
wall murals which are reported at some La Florida missions.
Blocks
A-B
Units 13, 28-29
Shell Midden
Total Percent
count
Percent
Wt
San Pedro 1273
7603.0g
107
859.7g
1380
8462.7g 58.1 59.3
San Marcos* 724
4503.8g
8
92.4g
732
4596.2g 30.8 32.2
Other 253
1166.5g
12
41.9
265
1208.4g 11.1 8.5
Total 2250
13273.3g
127
994.0g
2377
14267.3g 100 100
*Combines San Marcos, Irene, and colonowares
Table 9. Ceramic summary (count/weight in grams).
Excavations at the Armellino Site (8DU631)
Ashley, Rolland, and Thunen 71
Notes
1. San Pedro pottery is a grog-tempered ware manufactured
by the indigenous Mocama from about A.D. 1450 to 1600 (or
a little later). San Marcos, known as Altamaha in Georgia, is a
grit-tempered ware manufactured by the local Mocama during
the opening decades of the seventeenth century or the early
Mission period. Its production quickly supplanted that of San
Pedro by the Mocama. San Marcos was also made concurrently
by the Guale and Yamasee Indians of northeastern Florida and
southeastern Georgia.
2. Depth of all features was measured from midden-
subsoil interface, approximately 40 cm below surface, to
deepest point of the feature in prole.
3. Detailed zooarchaeological feature data are available at
the UNF Archaeology Lab.
Acknowledgments
Fieldwork on Big Talbot Island was partly funded by UNF
Summer Faculty Grants in 1998 and 1999. We are grateful
for all the help provided by UNF students and volunteers,
particularly Jim Wheat, Amy Bennett, David Nelson, Nate
Smith, Tracy Milligan, and Mike Foster. Special thanks to Dr.
Jeffry Will and Timothy Cheney of the Center for Community
Initiatives and Bob Joseph, former Manager of Big Talbot
Island State Park. We also appreciate the comments of two
anonymous reviewers and all the assistance provided by the
editors, Jeff Du Vernay and Julie Rogers Saccente.
References Cited
Ashley, Keith
2009 Straddling the Florida-Georgia State Line: Ceramic
Chronology of the St. Marys Region (A.D. 1400-
1700). In From Santa Elena to St. Augustine:
Indigenous Ceramic Variability (A.D. 1400-1700),
edited by Kathy Deagan and David Hurst Thomas,
pp. 125-140. Anthropological Papers, American
Museum of Natural History, Number 90.
2013 Grafting onto the Native Landscape: The Franciscan
Mission System in Northeastern Florida. In From
La Florida to La California: The Genesis and
Realization of Franciscan Evangelization in the
Spanish Borderlands, pp. 143-163, edited by
Timothy J. Johnson and Gert Melville. The Academy
of American Franciscan History, Berkeley, CA.
2014 Distribution of Contact and Early Mission Period Sites
in the Mocama Province. The Florida Anthropologist
67:159-178.
Ashley, Keith, and Rebecca D. Gorman
2011 Archaeological Survey of the Slough Sites, Fort
George Island, Florida. Report on le, Division of
Historical Resources, Tallahassee.
Ashley, Keith H., and Robert L. Thunen
2000 Archaeological Survey of the Southern One-Third of
Big Talbot Island, Florida. Report on le, Division of
Historical Resources, Tallahassee.
2008 Reexamining an Archaeological Survey of Big Talbot
Island. The Florida Anthropologist 61:133-148.
Ashley, Keith, Vicki Rolland, and Robert L. Thunen
2013a University of North Florida Archaeological Field
School 2009, 2011, 2012: Excavation of Block C
at the Cedar Point Site (8DU81). Report on le,
Archaeology Lab, University of North Florida,
Jacksonville.
2013b Missions San Buenaventura and Santa Cruz de
Guadalquini: Retreat from the Georgia Coast. In Life
Among the Tides: Recent Archaeology on the Georgia
Bight, pp. 395-422, edited by Victor Thompson and
David Hurst Thomas. American Museum of Natural
History Anthropological Papers No. 98.
Ashley, Keith, Robert L. Thunen, and Vicki Rolland
2010 Betz-Tiger Point and Cedar Point Preserve; Survey
and Field School, Phase II. Report on le, Division
of Historical Resources, Tallahassee.
2012 Field Season 3 (2011) at the Cedar Point West site
(8DU63). Report on le, Archaeology Laboratory,
University of North Florida, Jacksonville.
2013 A French Colony Lost: The Search for La Caroline
Project. Report on le, Division of Historical
Resources, Tallahassee.
Bennett, Charles E. (translator)
1975 Three Voyages: Rene Laudonniere. University
Presses of Florida, Gainesville.
Blanton, Dennis B.
2016 Mississippian Smoking Ritual in the Southern
Appalachian Region. University of Tennessee Press,
Knoxville.
Feest, Christian F.
1988 Jacques Le Moyne Minus Four. European Review of
Native American Studies 1(1)33-38.
Gibson, Guy
1998 Wall-Trench Structures. In Archaeology of Prehistoric
North American: An Encyclopedia, p. 870, ed. by
Guy Gibson. Garland Publishing Inc., New York.
Hann, John H.
1986 Translation of the Ecija Voyages of 1605 and 1609 and
the González Derrotero of 1609. Florida Archaeology
2:1-80, Florida Bureau of Archaeological Research,
Tallahassee.
1990 Summary Guide to Spanish Florida Missions
and Visitas with Churches in the Sixteenth and
Seventeenth Centuries. The Americas 46:417-513.
The Florida Anthropologist
72 2016 69 (1)
1996 A History of the Timucua Indians and Missions.
University Press of Florida, Gainesville.
Jones, William
1988 A Report on Big Talbot Island, Duval County, Florida
Ms. on le, Haydon Burns Library, Jacksonville.
Lafond, Arthur A., and Keith H. Ashley
1995 The Queen Mound (8DU110). The Florida
Anthropologist 48:35-45.
Lawson, Sarah (translator)
1992 A Foothold in Florida. The Eyewitness Account of
Four Voyages Made by the French to that Region
and Their Attempt at Colonization, 1562 1568, Based
on a New Translation of Laudonniere’s L’Histoire
Notable de la Florida. Antique Atlas Publications,
East Grinstead, West Sussex, England.
Keene, Deborah A., and Ervan G. Garrison
2013 A Survey of Irene Phase Architecture on the Georgia
Coast. In Life Among the Tides: Recent Archaeology
on the Georgia Bight, pp. 289-316, edited by Victor
Thompson and David Hurst Thomas. American
Museum of Natural History Anthropological Papers
No. 98.
López, Baltasar
1602 Letter to Blas de Montes, September 15, 1602.
Archivo General de Indias, Sevilla, Santo Domingo
235. Woodbury Lowery Collection, Library of
Congress, reel 2, Washington D.C. Translated by
John Hann.
Milanich, Jerald T.
1973 A Deptford Phase House Structure. The Florida
Anthropologist 26:105-118.
2005 The Devil in the Details. Archaeology (May/June):26-
31.
Nidy, Lynn S.
1980 Historical, Architectural and Archaeological Survey
of Duval County, Florida. Miscellaneous Project
Report Series Number 12. Division of Archives,
History and Records Management, Department of
State, Tallahassee.
Pareja, Francisco
1602 Letter to Blas de Montes, September 14, 1602.
Archivo General de Indias, Sevilla, Santo Domingo
235. Woodbury Lowery Collection, Library of
Congress, reel 2, Washington D.C. Translated by
John Hann.
Ribault, Jean
1964 The Whole and True Discouerye of Terra Florida.
Facsimile Reprint of the London Edition of 1563.
University of Florida Press, Gainesville.
Rolland, Vicki L., and Keith H. Ashley
2000 Beneath the Bell: A Study of Mission Period
Colonoware from Three Spanish Missions in
Northeastern Florida. The Florida Anthropologist
53:36-61.
Russo, Michael, Ann S. Cordell, and Donna L. Ruhl
1993 The Timucuan Ecological and Historical Preserve
Phase III Final Report. National Park Service,
Southeast Archeological Center, National Park
Service, Tallahassee.
Sturtevant, William C.
1992 The Sources for European Imagery of Native
Americas. In New World of Wonders: European
Images of the Americas, 1492-1700, ed. by Rachel
Doggett, 25-33. Folger Shakespeare Library,
Washington, D.C.
Thomas, David Hurst
2008 Native American Landscapes of St. Catherines Island,
Georgia: III. Synthesis and Implications. American
Museum of Natural History Anthropological Papers
No. 88.
73
Appendix A: Excavations at the Armellino Site Data
Ashley, Rolland, and Thunen
Block-
Unit
Plan size
NS x EW
Depth Description Contents Feature type
1A-4 28 x 32 cm 2.10-2.30
(20 cm)
Dark brown sand (10 YR 3/3) with some charcoal
ecking and shell fragments; sides slope slightly
inward with a round bottom (bowl shaped)
San Pedro 1/20.8g
Animal bone 3/2.5g pit
2A-4 30 x 40 cm 2.00-2.09
(9 cm)
Upper zone of very dark brown (10 YR 2/2) sand
with common crushed shell and charcoal ecks (6 cm
thick) followed by dark yellowish brown ( 10 YR 3/6
– 4/6) sand with rare charcoal (3 cm); bowl shaped
San Pedro 1/2.3g
Animal bone 2/4.1g
Iron oxide 0.9g
Corn 0.01g
pit
3A-4
A-9 30 x 18 cm 2.10-2.33 (23 cm)
Black core with common crushed shell (and whole
tagelus) and charcoal; lower section consisted of
mottled brown (10 YR 3/3) and yellowish brown (10
YR 4/6) sand with crushed shell; evidence of ther-
mal alteration to sand at interface between these two
zones; one steeply sloped and one slightly sloped side
and round bottom
San Pedro 3/29.8g
St. Marys 1/3.2g
Sand tempered 1/1.8g
Sherd < 2cm 4/5.4g
Iron oxide 95.2g
Animal bone 6/6.5g
Corn 0.02g
pit.
4A-4
A-8 26 x 36 cm 2.10-2.52 (42 cm)
Dark brown (10 YR 4/3) sand with frequent crushed
shell and small chunks and ecks of charcoal; straight
sides and round bottom
San Pedro 2/6.5g
Sand tempered 1/4,.6g
Sherd < 2cm
Iron oxide 1.3g
Animal bone 4/0.3
Corn 0.7g
post mold/hole
5A-4
A-8 33 x 34 cm 2.10-2.34 (24 cm)
Very dark grayish brown (10 YR 3/2) with crushed
shell and frequent charcoal ecking; slightly tapered,
parallel sides and round bottom
San Pedro 1/2.6g
St. Marys 1.2.5g
Sand tempered 7/31.5g
Sherd < 2cm 1/1.2g
Iron oxide 13.0g
pit
6A-11 28 x 42 cm 2.02-2.11 (9 cm) San Marcos sherd concentration (vessel) containing a
giant Atlantic cockle shell (small square perforation)
San Marcos sherds, modied
giant Atlantic cockle shell San Marcos vessel
7A-5
B-14 22 x 22 cm 2.10-2.23
(13 cm)
Very dark gray (10 YR 3/1) sand with some shell and
charcoal ecking; bowl shaped
San Pedro 2/14.2g
St. Marys 1/6.1g
Sherd < 2cm 5/9.9g
Animal bone 8/4.0g
Iron oxide 0.01g
pit
8A-6,7
B-16 25 x 30 cm 2.10-2.38 (28cm)
Black (10 YR 2/1) core with frequent crushed shell
surrounded by very dark grayish brown (10 YR 3/2)
sand with less crushed shell and charcoal ecks; rare
whole shell; bowl shaped
San Pedro 13/186.4g
Sherd < 2cm 8/11.4g
Iron oxide 22.1g
Corn 0.02g
pit
Appendix A
The Florida Anthropologist
74 2016 Vol. 69 (1)
Block-
Unit
Plan size
NS x EW
Depth Description Contents Feature type
9A-6 42 x 46 cm 2.05-2.36
(31 cm)
Dark brown sand (7.5 YR 3/2) with charcoal and shell
ecking; resembles basin shape to west but eastern
half is much deeper with tapered sides and somewhat
pointed bottom
Sand tempered 1/2.1g
Sherd < 2cm 2/1.7g
Iron oxide 0.01g
pit.
10 A-5
A-6 33 x 37 cm 2.10-2.29
(19cm)
Brown (10 YR 4/3) sand with some shell fragments;
rare charcoal ecks; bowl shaped with a wedge of
yellowish brown (10 YR 5/4) soil protruding into base
Sherd < 2cm 2/1.7g
Corn 0.01g pit
11
A-7 10 x 15 cm 2.10-2.57
(37 cm)
Very dark grayish brown (10 YR 3/2) sand with some
charcoal ecking and yellowish brown (10 YR 5/4)
mottling; sides taper to a pointed bottom
San Pedro 1/6.4g
sherd < 2cm. 1/1.3g
Iron oxide 0.01g
Corn 0.01g
postmold/hole
12 A-6 22 x 22 cm 2.10-2.33
(23 cm)
Very dark gray (10 YR 3/1) core with crushed shell
and charcoal ecking; slightly tapering sides and
round bottom
San Pedro 3/45.2g
Sherd < 2cm 4/2.1g
Iron oxide 0.8g
Corn 0.01g
postmold/hole
13 A-7 50 x 36 cm 2.02-2.29 (27 cm)
Very dark grayish brown (10 YR 3/2) core, with some
yellowish brown (10 YR 5/4) mottling and chunks
of charcoal and broken shell (some burned); bowl
shaped.
San Pedro 6/51.6g
Sherd < 2cm 5/10.9g
Quartz peeble 1/5.6g
Iron oxide 0.4g
Animal bone 5/9.8g
Corn 0.3g
pit
14 A-7
A-9 20 x 16 cm 2.10-2.35 (25 cm)
Very dark brown (7.5 YR 3/2) upper zone (5 cm
thick) with occasional crushed shell and rare charcoal
ecks followed by a poorly dened area of brown
(7.5 YR 4/3) sand with dark yellow brown (10 YR
4/6) mottles, rare charcoal ecks, and no shell; rather
amorphous shape with somewhat pointed base
Sand tempered 1/3.5g
Iron oxide 0.5g postmold/hole
15 A-8 40 x 35 cm 2.10-2.60 (50 cm)
Eastern half was bowl shaped (21 cm deep) and west-
ern half had classic post prole with straight sides and
rounded bottom (50 cm deep). Soil was very dark
gray (10 YR 3/1) sand with post having black (10 YR
2/1) mottling; charcoal ecking throughout both.
San Pedro 15/188.5g
Sherd < 2cm 18/27.0g
Iron oxide 9.3g
Animal bone 12/9.5g
Corn 0.8g
posthole and postmold
16 A-9
A-12 25 x 32 cm 2.10-2.40 (30 cm)
Upper part consists of dark brown (7.5 YR 3/2) sand
with common crushed shell and charcoal ecking
(rare charcoal chunks); lower part consists of same
dark brown with yellow brown (10 YR 4/6) mot-
tling and rare chunks of charcoal and shell ecking.
Straight, parallel sides with round base.
Corn 0.02g postmold/hole
75
Appendix A: Excavations at the Armellino Site Data
Ashley, Rolland, and Thunen
Block-
Unit
Plan size
NS x EW
Depth Description Contents Feature type
17 A-9 30 x 17 cm 2.10-2.22
(12 cm)
Very dark grayish brown (10 YR 3/2) soil with little
shell and charcoal ecking followed by the same
colored soil with dark yellow brown (10 YR 4/6)
mottles and no shell; rare charcoal; bowl shaped.
San Pedro 1/5.1g
Iron oxide 1.7g pit
18 A-9 46- 36 cm 2.10-2.24 (14 cm)
Dark brown (10 YR 3/3) sand with rare crushed shell
and frequent charcoal ecking; base consists of dark
brown (10 YR 4/3) sand mottled with dark yellow-
ish brown (10 YR 4/4) sand and no shell or charcoal;
bowl shaped
San Pedro 1/18.7g
Corn 0.01g pit
19 A-8 37 x 38 cm 2.10-2.30
(20 cm)
Dark brown (7.5 YR 3/2) sand with yellowish brown
(10 YR 5/4) mottling; charcoal ecking and crushed
shell is present (mostly stout tagelus and Atlantic
ribbed mussel; some larger but fragmented oyster);
bowl shaped
San Pedro Plain 5/46.2g
St. Marys 2/6.6g
Sherd < 2cm 8/17.1g
Spanish olive jar 1
Decortication Chert ake 1
Iron oxide 12.9g
Animal bone 2/0.5g
Corn 0.3g
pit
20 A-11
A-12 105 x 72 cm 2.10-2.17
(7 cm)
Core area of brown (10 YR 4/3) sand with rare broken
and crushed shell; outer area is amorphous, faint, and
mottled dark yellowish brown (10 YR 4/4)
San Pedro 2/22.1g
Sand tempered 1/7.4g
Sherd < 2cm 2/4.8g
Secondary chert ake 1
Iron oxide 4.8g
Corn 0.01g
Structure
entry?
21 A-10 20 x 24 cm 2.10-2.24
(14 cm)
Very dark grayish brown (10 YR 3/2) sand with yel-
low brown (10 YR 2/6) mottles; some charcoal ecks
and rare chunks; bowl shaped
Sand tempered 1/4.5g
Sherd < 2 cm 1/2.0g
Iron oxide 1.9g
Corn 0.2g
pit
22 A-10 22 x 38 cm 2.10-2.41
(31 cm)
Upper zone consists of very dark grayish brown (10
YR 3/2) sand with occasional shell and charcoal
ecking (includes area of dense charcoal); lower zone
is dark yellowish brown (10 YR 3/4 and 4/6) with
rare shell and charcoal; bowl shaped.
San Pedro 3/22.9g
Sherd < 2cm 9/15.9g
Iron oxide 17.0g
Decortication chert ake 1
Animal bone 6/2.7g
Corn 0.5g
pit
23 A-10 30 x 26 cm 2.10-2.45 (35 cm)
Very dark gray (10 YR 3/1) soil gives way to highly
mottled yellowish brown (10 YR 5/4) sand; large
piece of hematite encountered at 2.18 mbd; straight
parallel sides and round bottom
San Pedro 2/33.4g
Sherd <2 cm 6/11.4g
Iron oxide 250.8g
Animal bone 21/10.9g
Corn 0.4g
postmold/hole or pit
The Florida Anthropologist
76 2016 Vol. 69 (1)
Block-
Unit
Plan size
NS x EW
Depth Description Contents Feature type
24 A-6 20 x 20 cm 2.10-2.51 (41 cm)
Dark grayish brown (10 YR 4/2) sand with few yel-
lowish brown (10 YR 5/4) sand mottles; rare to infre-
quent charcoal ecks; straight sides and round bottom
San Pedro 1/27.3g. postmold/hole
25 A-5 18 x 18 cm 2.32-2.68 (36 cm)
revealed beneath Feature 7; grayish brown (10 YR
5/2) sand with brown (10 YR 5/3) and light brown (10
YR 6/4) mottling; slightly irregular sides that taper to
a point, but edges are not as distinct as drawn
Sherd < 2cm.1/1.0g postmold/hole
26 A-5 40 x 20 cm
(partial) 2.10-2.60 (50 cm)
Very dark gray (10 YR 3/1) soil with shell fragments
and charcoal ecks; only partly excavated because it
extends into east wall of Block A; bowl shaped
San Pedro 9/97.0g
Iron oxide 1/4.4g
Corn 0.01g
pit
27
A-9
A-12
B-21
25 x 25 cm
(partial)
2.10-2.63
(53 cm)
Dark grayish brown (10 YR 3/2) sand with scattered
crushed shell throughout; only partly excavated be-
cause it extends into south wall of Block B; parallel
sides and slightly rounded bottom
San Pedro 9/141.9g
Sand tempered 1/5.0g
Sherd < 2cm 2/4.3g
Iron oxide 1/0.7g
Corn 0.1g
postmold/hole
31 B-17 32 x 35 cm 2.00-2.58
(58 cm)
Mostly very dark brown (10 YR 2/2) soil; animal
burrow of dark yellowish brown (10 YR 4/4) ashy
soil with frequent charcoal ecking and chunks that
passes through the feature and continues beyond both
prole edges; straight sides and round bottom
San Pedro 5/95.5g
St. Marys 2/11.2g
Sand tempered 1/7.1g
Sherd < 2cm 12/18.8g
Iron Oxide 1.0g
Corn 0.3g
postmold/hole
32 B-17 23 x 32 cm 2.00-2.18
(18 cm)
Black (10 YR 2/1) soil with abundant charcoal and
occasional crushed shell; irregular edges; bowl shaped
Sand tempered 1/5.8g
Sherd < 2cm 7/6.9g
Carbonized corn cob 1
pit (re pit)
33 B-14 26 x 26 cm 2.23-2.63
(40 cm)
Formerly feature 7b; dark grayish brown (10 YR 4/2)
sand with some yellowish brown (10 YR 5/4) mot-
tling; straight sides and round bottom
Sherd < 2cm 1/0.g
Iron oxide 0.1g
Corn 0.01g
postmold/hole
34 B-15 24 x 28 cm 2.10-2.58
(48 cm)
Very dark brown (10 YR 2/2) soil with small patches
of ash and frequent crushed shell; irregular sides and
at bottom
San Pedro 2/18.7g
St. Marys 1/14.2g
Sherd < 2cm 12/20.5g
Iron oxide 0.8g
Corn 0.01g
postmold/hole
35 B-18 23 x 23 cm 2.00-2.07 (7 cm)
Yellowish brown (10 YR 5/4) sand with whole and
broken stout tagelus shells. Basin shaped with at
bottom
None pit
77
Appendix A: Excavations at the Armellino Site Data
Ashley, Rolland, and Thunen
Block-
Unit
Plan size
NS x EW
Depth Description Contents Feature type
36 B-14
B-15
223 x 161
cm 2.10-2.30 (20 cm)
Upper section is amorphous with very dark grayish
brown (10 YR 3/2) sand, moderate charcoal ecking
(some chunks), and scattered shell; it ranges in depth
from 5 to 30 cm with an undulating base; a 20 cm
diameter postmold (36a) is revealed from 2.30-2.58.
San Pedro 11/90.9g
St. Marys 1/4.1g
Sand tempered 1/3.7g
Grit tempered 1/10.0g
Sherd < 2cm 18/22.6g
Chert shatter 1
Greenstone celt frag 1/207.0g
Iron oxide 0.7g
Animal bone 53/5.8g
upper area of burning
(F-36) and postmold
(F-36a)
37 B-15
B-16
45 x 40 cm
(20 x 20 cm) 2.10-2.40 (30 cm)
Distinct area within Feature 36 that consists of a core
area of dark brown (10 YR 3/3) sand with sparse
shell and moderate charcoal ecking that thins and
becomes more mottled (yelowish brown, 10 YR 4/4)
with depth; base narrows to form a post prole (20
cm in diameter)
San Pedro 5/65.5g
St. Marys 1/3.6g
Sand tempered 1/2.3g
Sherd < 2 cm 7/15.2g
Bone sh hook 1
Animal bone 4/1.18g
Iron oxide 0.7g
Corn 0.01g plus 2 cobs
posthole and postmold
38 B-16 25 x 28 cm 2.11-2.38 (27 cm)
Very dark grayish brown (10 YR 3/2) sand with
sparse shell and some charcoal ecking (few large
pieces of charcoal); bowl shape
San Pedro 4/98.0g
Sand tempered 1/3.6g
Sherd <2cm 3/6.1g
Iron oxide 1.7g
Corn 0.5g
pit
39 B-25 52 x 31 cm 2.11-2.40
(29 cm)
Dark brown (10 YR 3/3) sand with charcoal ecks
and rare shell; one straight side with the other is
slightly angled inward, with a round bottom.
San Pedro 2/10.8g
St. Marys 1/2.7g
Sand tempered 1/2.1g
Sherd < cm 2/2.4g
Animal bone 1/0.1g.
Corn 0.01g
postmold/hole
*
40 B-25 37 x 24 cm 2.12-2.40 (28 cm)
Dark brown (10 YR 3/3) with some charcoal ecks;
roughly bowl shaped with western edge more dened
than eastern edge
San Pedro 2/14.2g
St. Marys 1/7.3g postmold/hole
41 B-16 11 x 6 cm 2.09-2.13
(4 cm)
Shallow concentration of charcoal that includes
charred corn cobs Corn 2.3g charcoal (corn) concen-
tration
42 B-19 34 x 34 cm 2.10-2.22
(12 cm)
Very dark grayish brown (10 YR 3/2) with some
yellowish brown (10 YR 5/6) sand mottles; some
charcoal ecking and very little crushed shell; bowl
shaped.
St. Marys 1/4.7g
Sand tempered 2/14.5g
Sherd < 2cm 2/3.0g
Iron oxide 0.01g
Corn 0.01g
pit
The Florida Anthropologist
78 2016 Vol. 69 (1)
Block-
Unit
Plan size
NS x EW
Depth Description Contents Feature type
43 B-19 33 x 25 cm 2.11-2.73
(62 cm)
Very dark grayish brown (10 YR 3/2) soil with some
shell fragments and charcoal ecking; straight sides
and round bottom
San Pedro 1/2.9g
Sand tempered 1/2.7g
Sherd < 2cm 5/6.2g
Animal bone 1/10.7g
Iron oxide 0.4g
Corn 0.01g
postmold/hole
44 B-26 17 x 22 cm 2.11-2.15
(4 cm)
Very dark grayish brown (10 YR 3/2) with yellowish
brown (10 YR 5/6) mottling; some shell pieces; shal-
low with bilobal base
Animal bone 1/0.2g postmold/hole
45 B-14
B-15 21 x 20 cm 2.18-2.33
(15 cm)
Identied within Feature 36; very dark brown (10
YR 2/2) with sparse shell and charcoal ecking; rare
charcoal chunks; slightly angled, parallel sides and a
round bottom.
Sherd < 2cm 1/2.7g postmold/hole
46 B-15 20 x 18 cm 2.12-2.67
(55 cm)
Soil is dark brown (10 YR 3/3) with yellow brown
(10 YR 5/6) mottling, common charcoal ecking, and
some shell; straight sides and round bottom
St. Marys 3/9.9g postmold/hole
47 B-24 29 x 17 cm 2.13-2.48 (35 cm)
Very dark grayish brown (10 YR 3/3) soil with
yellow brown (10 YR 5/8) mottling and some shell
fragments and charcoal ecking; round bottom with
straight sides (although one is partly collapsed)
San Pedro 1/9.0g
Sherd < 2cm 3/2.9g
Animal bone 1/0.3g
Whelk 1
Corn 0.01g
postmold/hole
48 B-24 30 x 30 cm 2.11-2.28 (17 cm)
Dark brown (10 YR 3/3) soil with brown (10 YR 5/3)
mottling; charcoal ecking and some shell fragments;
straight sides and round bottom, although upper sec-
tion is wider and slopes in
San Pedro 2/8.0g
St. Marys 1/6.0g
Sherd < 2cm 2/2.8g
postmold/hole
49 B-25
B-26 32 x 28 cm 2.13-2.45 (32 cm)
Dark brown (10 YR 3/3) sand with charcoal ecks
and some shell.; some yellow brown (10 YR 5/6)
mottling in lower part of feature; slightly tapered
sides with convex bottom
San Pedro 2/9.7g
Sherd < 2cm 1/1.3g
Animal bone 8/1.8g
Iron oxide 0.3g
Corn 0.1g
postmold/hole
50 B-22
B-23
65 x 65 cm
(20 x 20 cm) 2.10-2.40 (30 cm)
Dark grayish brown (10 YR 4/2) with some yellow
brown (10 YR 4/2) mottling; charcoal ecking and
some shell; less shell and charcoal in darker. core
area; irregular outer edges with post mold like core
(20 cm in diameter)
San Pedro 3/20.0g
Sand tempered 1/5.2g
Sherd < 2cm 3/6.1g
Animal bone 1/0.4g
Iron oxide 0.1g
Corn 0.01g
posthole and postmold
51 B-15 19 x 19 cm 2.10-2.13
(3 cm) Shallow charcoal concentration within Feature 36 . none charcoal concentration
79
Appendix A: Excavations at the Armellino Site Data
Ashley, Rolland, and Thunen
Block-
Unit
Plan size
NS x EW
Depth Description Contents Feature type
52 B-20 23 x 30 cm 2.10-2.27 (17 cm) Dark brown (10 YR 3/3) soil with yellow brown (10
YR 5/6) mottles and some shell; bowl shaped
San Pedro 1/9.9g
Sand tempered 1/3.4g
Animal bone 10/2.7g, corn
0.01g
postmold/hole