Certain conventions have been used in this study, in terms of both language and presentation. In general, I have tried to adhere to Fedele’s plea that “since we are dealing with ideology, being careful and neutral about terminology is a pre-requisite . . . caution is especially required in the use of terms which imply culture-bound overtones. And absolutely to be avoided are terms which in our own western cultures notoriously carry religious meaning” (Fedele 2004:52). I have, however, at times opted against this principle, most notably with my retention of the spiritually redolent word “amulet” to describe what might be brooches or pendants in the shape of animals and humans worn on the breast of certain stelae. On such occasions it is because the word has already been adopted in the literature on the topic and I judge it too late or potentially confusing to overturn the tradition. Several terms I have altered, but some I have elected to leave as is. All are flagged.
When referring to an element of the anthropomorphic imagery, something that is “worn” by the stele or constitutes a physical feature of its human makeup, I use left and right proper. That is, I speak in terms of the stele and not the viewer, as one would of the left ear of a statue or friend. When referring to the subsidiary figurative imagery upon the stelae, I do the opposite and refer to it from the point of view of the observer, discussing “the figure at left” as one would in describing a two-dimensional painting.
The term “double circle” describes a set of two concentric circles incised on the stele, “triple circle” a set of three concentric circles, and so on. The term “of uncertain gender” does not mean that the stele is of ambiguous sex, but that it is a fragment that retains no gender markers and cannot be categorized as either male or female.
The use of the word “stele” does not necessarily imply a complete stele but may also be used to refer to a single fragment or a collection of (usually) joining fragments from the same stele. For the purposes of statistics each unit is treated with equal weight, although it is acknowledged that this has the potential to skew results. A substantial number of new joins have been made; nonetheless it is not possible to say securely whether some (usually smaller) fragments do or do not belong to the same stele as other existing fragments, so that the same stele may be being counted two times or more. For the most part this is immaterial: a fragment from the hem is never going to carry a fibula and so, in any examination of the fibulae, it is irrelevant that an extant piece from the hem of the same stele is being treated separately in the same analysis. The hem fragment is simply omitted at the start as no fibulae can be expected to have been preserved upon it. The only statistics that can be truly biased by a fragmentary corpus are those that deal with them all, most notably analyses of the fabric designs used in the stelae’s robes and when making overall comparisons between the genders. It is, however, general trends that are of interest, and if the results are tilted a little to one side or the other it is of no grave consequence. It is the nature of archaeological data sets to be incomplete.
I have renumbered all the 1,400 stelae and stele fragments in this study according to my own system. This has been done to provide overall consistency, so that the stelae in the National Museum of Manfredonia logged by Maria Luisa Nava (Nava 1980a), the occasional finds from Ordona and the Melfese, those in the Trinitapoli Collection, pieces housed in museums in Geneva, Mariemont, Malibu, Budapest, and Kyoto, those held in private collections, stelae published long ago but now of unknown whereabouts, and stelae not published at all can all be dealt with smoothly and on equal terms. Juggling inventory numbers (or making my own arbitrary labels when there are none) proved to confuse the issue. The system is a simple numerical sequence beginning from 1, but divided by gender into three classes, each of which is denoted by a letter: F for female, M for male, and U for uncertain gender. The females number F1 to F695, the males M1 to M150, and those of uncertain gender U1 to U444. This has the great advantage of making the gender of any given fragment explicit without having to announce it. A fourth category exists for the heads, H1 to H111.
A concordance is provided (appendix 3), listing each stele’s Norman number (used in this publication), location, inventory number, and, where applicable, Stele Daunie I number from Nava 1980a. Those with multiple numbers from Nava 1980a have been joined since publication. The dimensions and typology classifications, as outlined by Nava (and summarized here in chapter 3, “Maria Luisa Nava”), of each piece are also given in this listing, as is their figure number, if illustrated. Elsewhere, if the Nava Type appears in parentheses, I have made the attribution myself. Appendix 2 provides information on stelae included in this study that have not been published in Nava 1980a or Nava 2011 (the Trinitapoli Collection). The spelling of “stele” and “stelae” I have adopted following the Oxford Editor’s Dictionary.
The term “Daunian” does not necessarily denote a discrete ethnic group, but instead describes the people living in Daunia at the time with a shared culture.