Griffin Mound Site (41UR142) Faunal Analyses

The investigation of the Griffin Mound site (41UR142) in the Little Cypress Creek basin in Upshur County, Texas, yielded 394 faunal specimens with a total assemblage weight of 127.71 grams. This sum includes all turtle shell, antler, and bone fragments. Faunal material was recovered from the site surface, four shovel tests, and four units in a 2 x 2 m excavation at this Middle Caddoan site, consisting of midden deposits and a large storage pit feature. The following sections of this article discuss the methods employed in the faunal analysis, results of taxonomic identification and quantification, and distribution of these remains.


INTRODUCTION
The investigation of the Griffin Mound site ( 41 UR 1 42) in the Little Cypress Creek basin in Upshur County, Texas, yielded 394 faunal specimens with a total assemblage weight of 127.71 grams. This sum includes all turtle shell, antler, and bone fragments. Faunal material was recovered from the site surface, four shovel tests, and four units in a 2 x 2 m excavation at this Middle Caddoan site (see Nelson ct al. 1996), consisting of midden deposits and a large storage pit feature. The following sections of this article discuss the methods employed in the faunal analysis, results of taxonomic identification and quantification, and distribution of these remains.

METHODOLOGY
All prehistoric vertebrate remains were inventoried and weighed, and Excel 5.0 for Windows was used to manipulate the generated faunal data. An Ohaus digital scale, Model CT600-S, was used to record bone weight. All fragments recovered were analyzed by the author, using comparative co11ections on loan from, or housed at, the Institute of Applied Sciences, Zooarchaeology Laboratory, University of North Texas, Denton, Texas. Occasional supplements were required, using conventional osteological keys such as Olsen ( 1964 ), Gilbert (1980), and Schmid (1972). Identifications were made to the most specific category possible depending on the condition of the bone and available comparative material. Only positive identifications resulted in the assignment of faunal elements to genus or species.
Standard zooarchaeological methods have been used, with first the animal bones inventoried and bagged by Archeological and Environmental Consultants (Austin and Pittsburg), then submitted to me for identification and quantification. Both unidentifiable and identifiable pieces were analyzed in similar fashion. That is, the same attributes were recorded: taxon, clement and portion of that element, anatomical location of the element, condition of the bone and any notes on age, taphonomy, burning or breakage patterns, and presence of modification, if applicable. Provenience information was also recorded.
Quantification of the assemblage is summarized as number of identified specimens per taxon (NISP) and as minimum number of individuals (MNI) for identified elements. MNI estimates were calculated according to the most frequently occurring element, based on symmetry and element portion (Munzel1986). In the mammalian class, teeth were used whenever possible.
The faunal data tables in this article are standard species lists with the number of occurrences for each animal. Those specimens regarded as unidentifiable (those coded to only class) have been consolidated into two general categories. Elements of non-diagnostic skeletal value (unidentifiable fragments and long bone shafts, see Olsen [l964]), are coded in an indeterminate category by class and size range. For example, specimens counted as "mammal" are from indeterminate-size mammals, and "large mammal" refers to a deer-size mammal. Recording these specimens in a size category enables the most precise level of observation as the specimen allows. In small samples, taking note of weight and the size categories of non-diagnostic elements broadens the function of the bone assemblage. However, percentages referred to in this report are calculated by NISP rather than weight.
Weights of specimens by Jot number can be found in the faunal data inventory, on file with Archeological and Environmenta.l Consultants.

RESULTS
The following section describes the vertebrate taxa recovered from the Griffin Mound site. Taxonomic classes identified include reptile, ave, and mammal (lagomorpha, rodentia, and artiodactyla). None of the faunal specimens arc modified. Number of identified specimens (NISP) and MNI for each taxon arc summarized in Table 1, as are weights for each taxon and percentages of site assemblage. The composition of anatomical elements can be found in Table 2.

Assemblage Composition Class Reptilia
Order Testudinata, Family Emydidae: Box turtle (Teffapene sp.) is represented by two shell fragmenls. One specimen was recovered from Feature 1 (an apparent storage pit feature, see Nelson et al. 11996 I) in Unit 5, LevelS (80-90 em bs). The second piece came from Level 4 (40-50 em bs) in Unit 7. Box turtles, which are strictly a North American species, range widely over the eastern and cenlral United States and into the Southwest, and they also occur in many parts of Mexico. These are dry-land turtles that close their shells tightly when danger threatens (Conant 1975). Both pieces arc burned.
Order Testudinata (family indeterminate): Twenty shell fragments from unidentifiable turtle were recovered from severctlleve1s (Levels 2-5 (20-60 em bsj and 7-10 [70-162 em bs I) of all four units, including six pieces from Feature 1, and one fragment found in Shovel Test 4. All of the specimens are burned.

Class Aves
Order indeterminate: An unidentifiable large bird is represented by one bone fragment recovered from Shovel Test I (0-20 em bs). This specimen compares favorably to turkey, but because of fragmentation, a specific identification was not possible. Turkey occurs as wild fowl in open woodland cnvironmenls (Robbins et al. 1983). This could be the remains of any one of a number of large game birds, but the Caddos were known to hunt turkey (Newcomb 1961).

Class Mammalia
Order Lagomorpha, Family Leporidae: Cottontail rabbit (Sylvilagus sp.) is represented by four specimens. They were recovered from Shovel Test 2 and Test Units 7, 8, and 617/8 in various levels, including one specimen from Feature I. Three fragments are burned. Currently, two species of cottontail inhabit this part of Northeast Texas: the Eastern cottontail (Sylvilagus jloridanus) prefers heavy brush, strips of forest with open areas, edges of swamps, and weed patches; swamp rabbit (Sylvilagus aquaticus) prefers swamps, marshes, and wet bottomlands (Burt and Grossenheider L 980). Osteologically, the swamp rabbit is the largest of the cottontails within its range ( Davis 1978 Teeth" inc. enamel fragments as well as complete teeth . Those in brackets were not included in site totals._ "Cranial" inc. skull eiements, mandible, and maxilia"Tr'agments. ..
-1--i ---Test Unit 8. The specimen is not burned. Preferring grassland, alfalfa fields, pastures. roadsides, and railroad rights-of-way (Burt and Grossenhcidcr 1980), this is the only pocket gopher found in the vicinity. It is probably a modern intrusion.
Order Artiodactyla, Family Cervidae: Deer (Odocoileus sp.) is represented by 35 elements, comprised of complete teeth and tooth enamel fragments. Seven enamel fragments arc burned. They were found in Shovel Test 4, and Test Units 5, 7, and 8. from Levels l through 8 (midden and feature contexts) as well as the surface. A minimum of one individual is represented. but age could not be determined because of fragmentation. White-tailed Deer (Odocoileus virginianus) is the only species that currently occupies the general area, are they are found in forests, swamps, and open brushy areas (Burt and Grossenheider 1980). Prehistorically, other species may have been present, but the elements found in this site assemblage are from smaller individuals. White-tailed deer are known for their small size, as compared to the larger mule deer of the western United States.
In addition to this quantity, the 18 medium artiodactyl bones most likely also represent deer. Although pronghorn is also categorized as a medium-sized artiodactyl, a specific identification cannot be made based on post-cranial bone fragments. But since no pronghorn tooth fragments were recovered, it is probably safe to assume that no pronghorn were present at the site. Both animals are similar in size, but pronghorn antelope are found in open prairies and sagebrush plains well to the west of the Little Cypress Creek basin (see Burt and Grossenheider 1980). Medium artiodactyl (deer) is probably also represented in the unidentifiable large mammal category (n=278). Large mammal bones were recovered from all levels of all squares, except for Test Unit 5, Level 10 (100-l 10 em bs), and Test Unit 7, Level 8 (80-1 00 em bs ). Totals range from 1-17, and most of these fragments are burned ( n=20 1).
Family Cervidae is also represented by two antler fragments. They were recovered from Test Unit 5, Level 7 (70-80 em bs), and Test Unit 8, Level 2 (20-30 em bs). Both fragments are burned. These too are probably deer remains rather than elk.

Indeterminate
Only four percent of the Griffin Mound faunal assemblage is recorded as indeterminate vertebrate (n=18). These bone fragments are indiscernible even at the class level.
A large portion of the assemblage was not identifiable to taxon, but was sorted into categories by size and class. Unidentifiable large mammal remains dominate the faunal sample (n=278, 70%). Indeterminate mammal (n=8) and small mammal (n=6) bone fragments are present, but occurrences are minimal. The single medium-sized mammal bone fragment compares favorably to fox, but a specific identification was not attempted due to its fragmentation and warping from burning. Red fox (Vulpes fulva) and gray fox ( Urocyon cinereoargenteus) are current residents of the project area, and both are similar in size. The red fox prefers a mixture of forest and open country, while the gray fox can be found in chaparral, open forests. and rimrock country (Burt and Grossenheider 1980). However, the red fox has been introduced historically for purposes of sport at several localities in eastern and central Texas ( Davis 1978).

Assemblage Condition
In general, the faunal material from the Griffin Mound site is highly fragmented. explaining the low Identifiability rate. Taphonomic patterns are absent on 313 specimens (fable 3). Surface observations include exfoliation and abrasion. Seventy percent of the site sample is burned (n=278), probably a result of trash disposal. Summary of burned specimens can be found in Table 4, and the distribution of these burned remains can be found in Table 5.
In addition to weathering, spiral fracturing was recorded during the faunal analysis. Spiral fractures are the result of impact on bone, such as striking it with a hammerstone or breaking it on an anvil. It is a common, expedient technique used in tool manufacturing, bone processing, and refuse disposal. UsuaiJy associated with large mammal long bones, spiral fracturing can also occur during trnmpling, carnivore gnawing, or any other severe impacts not necessarily associated with human activity. Ten specimens are recorded as spirally fractured (Table 6). The remainder of the large mammal sample is angularly fractured, suggesting that the bone was broken when it was dry, rather than while green and fresh. Perhaps after processing, it was broken into smaller pieces for disposal.
Scavenging activities are minimal. Carnivore gnawing was observed on five specimens, while rodent gnawing was nonexistent (Table 6).

Distribution
The distribution of faunal remains within the midden and Feature 1 is summarized in Table 7. The following section organizes the Griffin Mound faunal collection according to recovery by shovel tests and test units. Specific recovery by unit and level can be found in Tables 8 through I 1. Surface Collection and Shovel Tests 1 through 4 Sixteen specimens were collected from the surface of the site, comprised of large mammal, medium artiodactyl, and deer bone fragments. Four shovel tests yielded a total of 24 faunal specimens, with the majority recovered from 60 to 80 em bs in Shovel Test 4 (n= 12). These samples yielded a variety of animals: unidentifiable turtle, large bird, indeterminate mammal, large mammal, cottontail, medium artiodactyl, and deer. Twentythree pieces are burned.

Test Unit 5
Eleven levels in Test Unit 5 yielded 10 I faunal specimens, and over 55% of the sample is comprised of large mammal remains (n=56). The remainder of the sample consists of indeterminate vertebrate, turtle, box turtle, unidentifiable mammals, medium artiodactyl, deer, and cervid. Level 4 (40-50 em bs) yielded the majority of specimens (n=26), while only one fragment was found in Level 10 ( 100-1 10 em bs) in Feature 1. Seventy-seven fragments are burned, and the majority of these pieces (n=21) came from Level 8 (80-90 em bs) in Feature I.

Test Unit 7
Eight levels in Test Unit 7 yielded a total of 82 faunal specimens, and 78% of this sample is large mammal ( n==64 ). The taxonomic composition is similar to that of Test Unit 6, but box turtle, cottontail, and deer remains were also recovered. The majority of fragments carne from Level? (70-80 em bs), but Levels 2 (20-30 em bs) and 6 (60-70 em bs), also in the midden, yielded relatively high quantities also. Forty specimens from this unit are burned, with totals from each level ranging from two to nine.

Test Unit 8
Eight levels in Test Unit 8 yielded 67 faunal specimens. This sample is also dominated by large mammal bone fragments (n==55, 82%). The remainder is comprised of indeterminate vertebrate, turtle, cottontail, pocket gopher, medium artiodactyl, deer, and cervid remains, although quantities for each are minimal. Totals from each level range from two to 15 bones. Sixty-four percent of the unit sample is burned (n==43).

Composite Units
Eighteen faunal specimens were recovered from Level 10 in Test Unit 6/7/8, at the base of Feature 1 (I 20-162 em bs ), comprised of turtle, small and large mammal, cottontail, and medi um artiodactyl. An additional four specimens came from Level 9 ( 100-120 em bs) in Test Unit 7/8. All of these pieces are burned.

SUMMARY
The midden at the Griffin Mound site (41UR142) yielded the majority of faunal remains (n=317, 80% ). The storage pit feature had 77 fragments, including the only small mammal bones and the possible fox bone. Otherwise, faunal representation is similar in both areas. The distribution of fauna by level is also similar throughout the archeological deposit, with two exceptions. Level 4 had the highest recovery (n=73), while Level 11 (110-130 em bs in Unit 5) yielded only five specimens. The faunal collection from the Griffin Mound site indicates a high reliance on large game animals, supplemented with turtle, bird, and small animals such as cottontail. The sample is very typical Caddoan subsistence refuse.