1Jean-Pierre Bocquet-Appel’s (JPBA) publications have, for the most part, achieved a significant readership. According to Google Scholar, the most cited article is undoubtedly “Farewell to paleodemography” (Bocquet-Appel and Masset 1982), with 661 citations1. Of all his publications tracked via Google Scholar, the average citation is around 56, with 34% of publications cited less than ten times. As with any career, the publication strategy depends on various linguistic, institutional, or financial constraints, which inevitably leads to a certain proportion of seldom or never cited publications. This rather dry statistic obviously reveals nothing about the intrinsic quality or interest of such publications. They are often the result of original work carried out on the fringe of the researcher’s dominant themes and may have been the subject of publications, often mediocrely edited in obscure journals or poorly promoted by the authors themselves. Some works fade away with time, quickly eclipsed by more visible, better published, or more relevant novelties. Is this wretched fate always deserved? Do such neglected works deserve further attention or are they merely the froth of ‘real’ science that dissipates over time?

2To illustrate these introductory remarks, I propose to review some of JPBA’s publications to highlight those that have not had the chance to acquire the readership they undoubtedly deserved. Were these works really irrelevant or too technical, or only published in inadequate journals? Or are they just an indicator of future work promised to gain a better audience by choosing a better publishing strategy? There are many explanations, as illustrated by the following examples.

3Consider the article on “Estimation de la stature de la classe féodale d’après les armures du xvi^e s.2”. This original work proposes to compensate for the lack of information on the elites’ stature of the time by measuring their heavy protective equipment. Despite its novelty, this article, published in “Ethnologie française,” a journal that claims to be a “reference in French-speaking social science,” is cited only twice. The first time, briefly, in a book by C. Theureau on the skeleton’s anthropology of the Saint Pierre-le-Puellier cemetery (Theureau 1985). The second citation is included in an article published six years after JPBA’s in the “Zeitschrift für Morphologie und Anthropologie” (Wurm 1985). The author uses the same methodology advocated by JPBA, this time applying it to the German Empire’s soldiers in the 16th and 17th centuries, JPBA data published in the “Ethnologie française” are abundantly reproduced and compared with those collected in Germany. Although this German publication is a testimony to the creativity of JPBA’s idea, it does not seem to have been followed up. Was it because the choice to publish his initial article in a French journal was an inappropriate one? Or should we judge that the initial idea of inferring the stature of 16th c. elites by the size of their armor, while both original and compelling, proved to be too anecdotal to have a future rich in citations?

4Another example is given by the article published more than 30 years ago, in collaboration with the author of this note (Darlu and Bocquet-Appel 1987). Its ambition was to infer the degree of kinship between two people based only on anthropometric data measured on the skeleton or the living. At the time, JPBA had just published his dissertation and then his book on the biological variation of the Portuguese population in the 19th c. (Bocquet-Appel and Xavier de Morais 1987), in which anthropometric measurements on skeletons featured prominently, while at the same time being exploited by multivariate methods and so-called distance analysis, which were very fashionable in the 1980s. For my part, as his coauthor, I was interested in Bayesian methods of kinship distance based on genetic data such as blood groups (Darlu and Cavalli-Sforza 1985)3. This potential synergy gave rise to this collaborative work, in which anthropometric measurements, reduced to their independent components after multivariate analysis, were integrated into a Bayesian model for expressing the degrees of relationship. Simulations were used to verify the effectiveness of the method. Published in the Bulletins et Mémoires de la Société d’Anthropologie de Paris, with quite a few typos, this article was cited only once, probably unread, in a dissertation by Charisse Carver (Carver 2015) for a PhD from the University of Arizona. The citation came only as evidence of the historical interest of French physical anthropologists in kinship research in the 1960s.

5Despite the interest of this Bayesian inference method, which was almost the only one at the time able to propose such a diagnosis of kinship based on physical measurements, it did not have the impact it could have had. Was the journal inadequate? The authors’ obvious lack of commitment to promoting this work in a publication with better visibility is probably the first reason. This paucity of citation can probably also be explained by the fact that this approach would, at the turn of the 1990s, quickly become totally obsolete with the arrival of DNA. Skeletal DNA sampling and identical-by-descent DNA length comparisons (Huff et al. 2011) would prove to be a much better method of determining kinship, to the extent that such kinship diagnostics are now being marketed by various commercial companies, which will not be advertised here.

6A third example is also instructive. JPBA and Lucienne Jakobi published their first article on the transmission of fertility, in French, in the journal Population (Bocquet-Appel and Jakobi 1993a). This article is based on genealogical data collected by Lucienne Jakobi in the village of Arthez-d’Asson (Pyrénées-Atlantiques) from the 18th to the beginning of the 20th c. JPBA applied the methods of path analysis widely used at the time in the field of epidemiological genetics (Li 1975). Although published in a good journal, this article has only been cited once in a general article on “Biologie et histoire de la population4” by Guy Brunet, in the Annales de démographie historique (Brunet 2004). In the article, Brunet refers to Bocquet Appel and Jakobi to illustrate “the impossibility of demonstrating whether there is a heritability of fertility” in a discussion raising the unrealistic nature of a simplistic dichotomy between biology and culture. The same year, a similar article was published by the same authors in the Annals of Human Biology (Bocquet-Appel and Jakobi 1993b). This one is cited 14 times! However, three of these 14 articles are authored by the same Michael Murphy (e.g. Murphy 2007), who does not fail to insert this reference, and not the one from Population, when he compares the heritability values he estimated and those reported by JPBA. This example shows how preferable it is to favor English-language journals over French-language ones. Yet, the French INED journal Population has a very good international audience and (now) has an English version.

7Let us now examine JPBA and Bacro’s article published in 1994 in the excellent journal Systematic Biology (Bocquet-Appel and Bacro 1994). They present a generalization of the Wombling model developed by Womble (Womble 1951) and Barbujani and colleagues (Barbujani et al. 1989). The article describes abrupt changes in a variable’s value in geographical space, such as an allelic frequency or a name, or a continuous variable such as stature or fecundity. The generalization proposed by JPBA considers several variables simultaneously while also factoring in the correlations that may link the variables together. This rather technical article seems to have been a real success since it is cited 45 times, although much less than the “Farewell to paleodemography,” no doubt because it is both more technical and above all because it is not controversial, which, as is well known, generates a multitude of citations. However, what about the specifics of these 45 citations? The vast majority of the citations are included in the articles only to point out the existence of Wombling’s methods, citing the three ‘classic’ authors on the subject (Womble, Barbujani, Oden), to which JPBA reference has now been added to form a quartet regularly cited as essential references for these statistical methods applied to geography. The various authors include these four citations mainly to demonstrate that their own methodological developments are far more innovative than those proposed by the former... Furthermore, very few citations mention the implementation of this generalized Wombling method. In fact, these citations are only those proposed by JPBA or Bacro themselves (Bocquet-Appel et al. 2002 ; Balabdaoui et al. 2001; Bocquet-Appel and Jakobi 1997), except for one (Darlu 2004), which was poorly used. In other words, JPBA has found the ability and the energy to make substantial improvements to the “Wombling” method and to write a complex program (in Fortran) to apply it to various types of data. Unfortunately, almost no one has ever used it except himself, Bacro, and Darlu through publications that are themselves seldom cited. If JPBA is rightly included in the history of Wombling, it is only for having made a brief incursion into it, which did not produce as many results as it could have. No doubt attracted by other themes dear to him, such as palaeodemography and settlement histories, JPBA thus abandoned this fine foray into the field of spatial description without having brought it to fruition.

8These few examples illustrate the atypical and evolving career of JPBA, with its varying publication strategies, which can lead, depending on the situation, to excellent articles in small minor journals and more opportunistic articles in wide-readership journals. These forgotten publications, erased by time, scarcely cited or cited by convention, thus reveal other facets of JPBA’s talent than those manifested by his successful articles. All of his work, both large and small, is a beautiful illustration of the fluctuating complexity of a scientific trajectory when guided by creative imagination and scientific rigor.