order to reduce the effects of these fluctuations, the number of placements in police custody per month has been calculated over a two-year rather than a one-year period in the table above. The populations of the districts are very different, as are the locations of the gendarmeries inspected. The unit of Sainte-Ménehould corresponds to a grouping of gendarmeries whose surface area extends over three cantons. Varennes-en-Argonne is a “local gendarmerie” whose authority necessarily covers a smaller district. Saint-Chéron has the characteristics of a peri-urban area with a considerably higher population. Finally, Tarascon-sur-Ariège is a gendarmerie with authority over large district, characteristic of the French department in which it is located. The comments made supra with regard to the national police, concerning the relation between crime and the local population, also apply for the above table, although to a lesser degree. The offences are more usually committed by inhabitants of the district, but in proportions difficult to determine and variable according to the case. Saint-Chéron, which is close to major axes of communication, shows offences committed in particular by persons from “outside”. This is not the case for Varennes (in spite of the close proximity of the A4 motorway). The higher proportion of minors in the placements in police custody at Sainte-Ménehould is explained by the presence of a young offenders’ institution in the commune: the share of crime attributable to the latter institution is in the region of 28% for the year studied (but for a limited volume of offences). There the comparison finds its limits. Finally, a table concerning fundamental rights, comparable to that for national police establishments, has not been drawn up, since the selected items are less significant in the case of gendarmerie units: thus, there are never showers in gendarmerie facilities49; the question of their use is therefore not an issue. Generally speaking, police custody facilities in gendarmeries are more basic, from the point of view of both bathroom facilities and available premises (rooms reserved for consultations with doctors and lawyers, or for noting particulars, are extremely rare). Two useful circumstances have nevertheless been singled out: on the one hand, the cleaning of secure rooms and, on the other hand, the systematic taking away of brassières from women placed in police custody. Having made these preliminary observations, what comments can be made about the table? Out of the four gendarmeries studied here, the highest rate of persons implicated is smaller than the lowest rate observed in the police stations (Douai). For the latter, the average observed rate is 29.3°/°° (23.6°/°° if Saint-Ouen and Chessy are excluded); for the gendarmeries here examined, it is 9.1°/°°, thus being three times lower. This very considerable difference does not tell us anything about crime in itself, but reflects differences attributable both to crime and to the practices of the police forces and gendarmeries in question. The differences are of the same proportions with regard to police custody, and even somewhat higher since the rate at which persons implicated are placed in police custody by gendarmes is in general lower than that resulting from police activity (35.6 % on average for the former; 60.4% for the latter)50. The proportion of minors in police custody is also lower, though not markedly so: 16.5 in “police zones”, 15.1 % for gendarmeries. However, the characteristics of police custody at SainteMénehould push up the rate in “gendarmerie zones” (ZGN [Zone Gendarmerie Nationale]). Similarly, the percentage of placements it police custody of more than 24 hours is lower in the gendarmeries (9.6 %) than in the police stations (13.9%): it may be tempting to conclude that this 49 The Contrôle Général has already pointed out that, even in the architectural projects for gendarmeries currently decided by the central administration, no provision is made for showers, which is highly regrettable. 50 It is recalled that, excluding traffic-related offences, the national PI/PC rate (persons implicated/police custody) was 49.3% in 2009 and 45.6% en 2010. 47