This text is concerned with religious practice: ritual process and spirit possession among Réunionnais of African and Malagasy descent. Françoise Dumas-Champion has been working in Réunion for the best part of two decades and the depth of her ethnography – reflecting such a long engagement – is impressive. Her intent here is to identify the sources of cultural, and specifically religious, practices in Réunion, and to explore the interactions between them and between adherents of different religious systems. As she points out in the conclusion, these interactions are nuanced by the fact that transformations in religious thought – such as the monotheistic character of local Malbar (Hindu) perceptions of deity – have tended to shift the different religions into the same register. The result is that appeals to Malagasy, Hindu or Catholic practice are all equally possible, acceptable and widespread.

Much of the text is devoted to fine-grained ethnography: detailed descriptions of rituals undertaken by Réunionnais confronted with a range of strategic choices, strategies aimed at curing illnesses or misfortune, re-establishing family relationships or invoking benedictions. However, the extensive detail inevitably leads to much repetition, and more analysis would have been welcome, and particularly a contextualization within wider society. Not until the penultimate paragraph of the book does Dumas-Champion refer to neighbouring islands, observing that, given the political conditions, such practices are largely absent in Mauritius and Seychelles: this raises the obvious question of what these political conditions might be. Of what relevance are local social tensions, extremes of wealth and poverty, unemployment rates that hover around 30 per cent and latent hostility towards privileged metropolitan French residents on the island? Similarly, although there were numerous scattered references (often in footnotes) to individual links with Madagascar, these were not elaborated upon: having read the book, I still have little idea of whether Réunionnais have maintained family links with Madagascar or whether social ties were ruptured in the trauma of forced migrations. Where do those Réunionnais who go to Madagascar go? Are they ‘reinventing’ social relationships with people (and practices) in areas with which they have no prior connections? That people of presumed Antanosy descent should appropriate Merina practices suggests that the answer to the latter question is yes; if so, then how are these practices appropriated, and what then is the relevance of Malagasy identity in wider spheres of social interaction?

One of the stated objects of this text is to prove that the origins of many cultural practices in Réunion lie in Madagascar and Africa; since no one (presumably) would deny such origins, the author sets herself the task of identifying these links in some detail. I found this aim problematic and its execution unconvincing. Sentences that begin ‘In the south of India, among the Kallar …’ (p. 75) conjure up the spectre of Sir James Frazer, and the somewhat eclectic and apparently speculative nature of Dumas-Champion's drawing on other cultural practices is unsustained and, I suspect, unsustainable. Conclusions drawn from the existence of a cultural practice in two different places should be treated with circumspection. Take photographs 3 and 4, for example: they depict a near-identical ritual pouring of flour into a cone, the former among the Yao, the latter from Réunion. The caption to photograph 4 states that this practice ‘est d'origine Yao … comme le prouve la photo no. 3’. It, of course, proves nothing of the sort, and this example is only one of many such ‘proofs’ of cultural origins.

But even if these links were proved, what of it? While it is not the author's intention to provide an historical analysis of movements within the region, there is no sustained investigation of the relevance of these links for the actors themselves: for while they frequently invoke generic Malagasy origins, there appears to be an almost complete absence of local actors' appeals to Bara (or Antanosy, or Yao) heritage for their practice. Indeed, on several occasions the author follows these ‘proofs’ of origin with statements such as ‘on ne connaît plus le sens de cette pratique’ (p. 145), that ‘la pratique est totalement assimilée sans qu'on se soucie de son origine’ (p. 250), or even that borrowings are totally denied (p. 251). It seems unlikely that anyone would enact practices without some sense of their meanings, and the almost complete absence of local exegesis of ritual practice and the production of identity was troubling.

Finally, there were some minor errors: goat is not and never has been a prohibited food to Muslims; and the use of the term ‘Bantu’ is problematic. ‘Bantu’ is not an ethnic group but a linguistic group, and while many Bantu-speaking peoples share certain cultural and social traits, many others do not. Thus, in many Bantu languages (including Comorian and Gikuyu), koko means grandmother, not grandfather. I also found the excessive number of footnotes irritating: many, possibly most, were repetitive or superfluous.

Although several of the questions raised by the text were acknowledged in the conclusion, they were considered with a regrettable brevity and inadequately analysed; the lack of contextualization within discourses on creolization was particularly frustrating. Nevertheless, the tracing of practice through the thread of individual trajectories illustrates very effectively how religion is strategically activated in a domestic context.