Seldom has the title of a book been more accurately descriptive. MacKinnon’s (henceforth M’s) goal is to assemble as complete a picture as possible of how animals were raised and used in Roman Italy. To accomplish this goal, he has assembled a database of the faunal remains (primarily bones and teeth) from 97 Italian sites, most of them dating in the period from the second century BC to the third century AD. He also summarizes what we learn from ancient textual sources, especially but by no means exclusively the writers on agriculture. The result is a composite — in M’s terms, an “integrated” — picture of Italian husbandry and the Roman diet. The book does not advance our understanding of Roman animal husbandry, nor of the Roman diet, in any major way, but it provides many interesting details and it is important as an experiment in method, for it illustrates clearly the potential inherent in the study of faunal remains as well as the problems involved in using zooarchaeological and textual evidence.
The book is organized in a rigorously logical fashion. After a brief introduction, M surveys the kinds of written texts in which information about animal husbandry and the consumption of animals can be found. “Production” in his title includes the raising and slaughtering of animals, but not the market mechanisms by which they were made available. Similarly, “consumption” means for M the use of animals for food, although he is well aware of their other uses, as draft animals or as sources of wool, leather, and other products (p. 11).
In Chapter 2, M gives a useful summary of the relatively new field of zooarchaeology and of the problems involved in using faunal remains. These include variations in the method of recovery (manually, say, or by wet or dry sieving); differences in the post-depositional history of the bones;1 methods of determining the age and sex of the animal at death; and, most fundamentally, methods of counting the surviving bones and making sense of the counts. M introduces the two most widely used systems of quantification: simply counting the “number of identifiable specimens” (NISP number); and calculating the “minimum number of individuals” (MNI number). While fully aware of the scholarly debate concerning these methods of quantification, M believes that when taken together, “NISP and MNI generally provide a good indication of the abundance of species or taxa” (p. 23).2
In Chapter 3, M begins to set out his basic data in tables and figures: the total sample, NISP sample, and NISP numbers for mammals, birds, fish, reptiles, and amphibians, site by site (Table 11); the figures for mammals alone (Table 14); the evidence on the state of preservation and methods of recovery (Table 9). Noting that practices and preferences may have varied across time and space, he subdivides his material, first by type of site (rural, large urban, small urban, and “special,” such as votive deposits), and secondly by geographical groups — north, central, and south Italy — and chronological periods, “Republican” (roughly 500-50 BC “Imperial” (50 BC-AD 300), and “Late Antique” (AD 300-500). All of this, and basic bibliography, is set out in Tables 1-8.
One is struck immediately by M’s industry. As M puts it, “compiling the data was challenging” (p. 32). In fact, it must have involved an enormous amount of work, endless decisions and compromises, and painstaking care, as a check of M’s figures for thirty of the sites in Table 11 revealed. Many of the archaeological reports are in obscure journals, and a significant number are still unpublished. The type and amount of information given varies enormously from report to report. Some include statistics on teeth, or on sex, or age, while others do not. Some include information only on mammals, others on birds, fish, and mollusks. Mammals are usually reported together and separately from birds, but sometimes all domestic animals, including chickens, are lumped together in opposition to wild animals. Faced with this welter of material, M worked through the problems and made his decisions very carefully.
The figures I found varied from M’s by one or two in a few cases, probably just as often my mistake in addition as M’s. In only one instance, the “total sample” size at Settefinestre Period 2, did our figures vary significantly (M gives 6742, while my figure was 8911 plus unidentified fish bones), and that particular figure is not one that affects M’s analysis. There are a few inconsistencies in M’s use of his chronological periods. At Pistoia, he assigns the excavators’ phase V (second century AD) to his own Period 3 (Late Antique), while at Naples, Carminiello he puts the excavators’ phase IVA (second century) in his Period 2 (Imperial), while Phase III (late first century AD) goes in his Period 1 (Republican). Occasionally, M’s source is not clear. Table 11, p. 55, contains an entry with a full set of statistics for Otranto in Period 1, but Table 1 (where one should find a reference to the archaeological report) has no mention of Otranto. Such omissions and inconsistencies are rare, however, and unlikely to have affected M’s conclusions in any significant way.
The four chapters that follow are the heart of the book. They deal with cattle (Chapter 4), sheep and goats (5), pigs (6), and butchery, meat preservation, and cooking (7). Each chapter consists of three parts. First, M sets out and analyzes, in relentless detail, the zooarchaeological evidence. Second, he assembles a composite picture of ancient practices and preferences as they emerge from written texts. The third section is an “integration” of the zooarchaeological and textual evidence, in which M discusses where the two types of evidence supplement one another, agree, or conflict. My comments will deal with these three types of material.
Zooarchaeology: The three chapters on specific animals are subdivided into topics that will be familiar to specialists: identification and quantification by NISP and MNI frequencies; age, based on epiphysial fusion and dental data; seasonality (or time of year when animals are slaughtered); sex; age and sex correlation; size; and abnormalities in the skeleton as indicators of disease. M presents his data in 103 tables, figures, and appendices accompanied by systematic and rigorous discussions of the possible biases and limitations of the evidence. His goal is to discover patterns or trends over time and space (p 42), and he qualifies most of the patterns he does note with words and phrases such as “in general,” “could be,” and “possible.”
M is quite right to formulate his conclusions in this tentative way. While his data are drawn from a total of 97 sites, many of them provide no evidence on some of the questions he is asking. Moreover, when we subdivide the sample into three geographical groupings and then again into four types of site, sample sizes are often very small. Table 23, for example, gives the data on correlation of age and sex of cattle. As M notes in his discussion (p. 83), only 23 sites provide relevant usable data here, “too few to draw any significant conclusions.” The sample size of cattle whose sex could be determined is ten or more at only six of the sites, and at six others it is just one or two. Table 44 provides evidence on the size of pigs. From all of southern Italy, the total sample is 15 specimens, or about 3 pigs per century. Our statistics, that is, are subject to change, and conclusions drawn from them must be, as M’s are, tentative. Sometimes the data conflict: “there is clearly much inconsistency between fusion and dental data” from sheep and goats (p. 107). In short, the material currently available will not allow many confident assertions concerning Roman husbandry or diet. It does, though, provide an excellent testing ground for zooarchaeological evidence and method.
Since this study does serve as such a test case, I was disappointed in M’s discussion of the relative merits of NISP and MNI figures. As he notes, the two sets of figures are in general agreement. But there are anomalies, which undermine one’s confidence in the validity of the figures, and a discussion based on specific examples but providing general guidance is needed. On pages 72-74, for example, the NISP and MNI frequency figures for mammalian taxa are set out. In Table 16a we find that the mean NISP frequency of cattle on special sites is 23.9% ( higher than on urban1 sites). From Table 17a, though, we learn that the MNI frequency of those same cattle is 12.1%, half the NISP frequency and considerably lower than on urban1 sites. Similarly, on rural sites, wild animals and rodents taken together have an NISP frequency of 6%, but an MNI figure of 17.7%, nearly three times as high. What do these variations mean? Should we be worried about them? This might have been a good occasion (and there are many others) for M to provide a general discussion of such discrepancies and their significance. Textual Sources: M as usual is systematic and reliable in presenting the written evidence.3 His composite pictures of Roman practices are logically organized, and he avoids the unusual and sensational. My question regarding these sections is simple: are they needed at all? Similar summaries are available in K.D. White’s Roman Farming, and, more importantly, M himself will repeat virtually all of the relevant points in his “integrating” section.4 Why not simply organize the “integrating” sections around subjects on which zooarchaeological evidence supplements, supports, or contradicts the textual evidence?
Integrating the Zooarchaeological and Textual Evidence: This is, of course, a logical thing to do, and it can be very instructive. The most interesting contributions M notes concern beef. In our texts, pigs are mentioned as food more often than cattle, and their NISP frequency counts are higher as well. But M, following a method employed by earlier scholars, notes that if the NISP and MNI figures are converted to “meat weight,” it turns out that Romans ate about as much beef as they did pork (pp. 193-94).
Yet there are problems that might be considered. Archaeological evidence can be readily divided into chronological periods. The textual material cannot, since one author might copy from another, of two centuries or more before. The two types of evidence agree and disagree in unpredictable and disconcerting ways. For example, they seem to agree on the size of northern cattle, but disagree on the size of Campanian cattle (p. 92); there are similar inconsistencies regarding sheep sizes (p. 123), and on pig-producing zones (p. 153). M works hard to explain each of these particular cases and many others, but it is really a more general problem, and we need a more general explanation of the relation of text to faunal remains. Sometimes there is a nearly total disconnect between the two. M spends a great deal of space on transhumance (pp. 113-14 and 125-29), but the currently available zooarchaeological evidence contributes almost nothing to our understanding of the practice. Similarly, there is discussion on p. 226 of the diet of the urban poor. It is not clear to me that this is needed: it is speculative and cannot be supported by zooarchaeological evidence. A shorter, more focussed, discussion might well be more helpful: what specific practices can zooarchaeology tell us about, and what do we learn from them when they are taken in conjunction with the textual evidence?
Conclusions: M deals in this book with just four animals, primarily as sources of meat, although he often provides statistics on other taxa as well. There are good reasons for his decision, but it has serious consequences. We are not learning much about the Roman diet as a whole, but about the preferences over time for one or another of the four animals. The historical conclusions that can be drawn are in general unsurprising. A few examples: bovines generally lived beyond three years (p. 80; they could be used for several years as draft animals); adult bulls were rare (p. 80; they are less tractable than cows or castrated males); the Romans did recognize and raise two or more breeds of these animals (cattle, p. 85; pigs, pp. 154-55; note M’s statement on p. 243, “perhaps the greatest value of integrating zooarchaeological and textual data comes in terms of the analysis of animal breeds and the recognition of breeding improvements”; but note also that even this evidence has not yet helped us identify the specific breeds); many and perhaps most animals were butchered outside of urban areas, the meat then being brought in as whole or partial carcasses (despite the lack of refrigeration), p. 184; cattle were an important source of meat, pp. 189-94; and in general meat was desired as a part of the diet, p. 217. These are all interesting points, but they make a rather modest contribution to our knowledge of ancient Italy.
Faunal remains have been much studied in recent years, and M makes good use of the work of earlier scholars for his methodology. Thus his methods of analysis in the zooarchaeological sections are not new. But for the most part bones have been studied on a site-by-site basis, and M’s regional approach is largely new, as he justifiably notes (p. 10). That, together with M’s systematic use of textual evidence, is the real contribution of this book. M has had to confront problems in the presentation and analysis of evidence that previous scholars have not grappled with. In so doing, he has begun to establish a methodology for such studies, he has resolved some of the problems, and he notes that further work will be much more reliable if zooarchaeologists adopt and systematically apply standard methods in the recovery, description, and reporting of zooarchaeological evidence.5 Such standardization would go far toward improving the utility of this promising type of evidence, and it is to be hoped that M’s appeal to his colleagues will be heeded.
1. Taphonomy, or the study of what happens to bones after they are discarded, has become a field in itself. M. summarizes it nicely and directs the reader to various studies, including that of R. Lee Lyman, Vertebrate Taphonomy (Cambridge 1994).
2. For a more thorough presentation of the debate than M has space for, and a negative view of the value of MNI figures, see Terry O’Connor, The Archaeology of Animal Bones (Stroud, UK, and College Station, TX, 2000) 54-67.
3. I checked footnotes 1-50 on pages 113-117. The level of accuracy and precision is not quite as high as in the zooarchaeological sections, but the errors I noted are mostly trivial. Cato, cited in n. 8, does not mention the spread of disease. Note 13: Varro does not mention Heraclea nor his own flocks in these passages. Is Rust. 2.2.9 meant? Note 25: no Buck 1983 appears in the Bibliography. Note 27: what is ERC? (A list of abbreviations is needed.) Note 31: I don’t find a “part 4” in the Loeb or Budé editions of Sidonius, and Sidonius Carm. 7.364ff. does not seem to mention seasons.
4. One example, of many such repetitions: “Following slaughter, pigs are normally dehaired or skinned” (p. 170); “Following slaughter and draining of blood, pigs are normally dehaired or skinned” (p. 178).
5. These would include: clear statements on recovery techniques including such matters as the mesh size of screens (p. 45); the development of a standard scale for reporting bone fragmentation (p. 51); the use of standardized statistics on bone size (p. 84 n. 7); and standards for reporting evidence on butchery (p. 163) and estimates of meat weight (p. 189). See further M’s own peroration and list of desiderata, pp. 248-49.