Material and methods
Data acquisitionnext section
We surveyed the wildlife markets in 20 towns throughout Morocco in April to June 2013, April to May 2014, and December 2014, as well as the Spanish exclaves of Ceuta and Melilla in northern Morocco in April 2014 (Fig. 1). We selected the towns to include all of the larger ones (including the capital Rabat), important trade hubs (including ferry ports such as Tangiers), and two border towns. We visited 13 markets only once, but surveyed others up to six times over all three survey periods. The total survey effort was 49 visits. We identified and counted reptiles sold for medicinal purposes, and we gathered data on usages and prices through informal conversations with vendors and medicine dispensers. In towns that we surveyed only once, we selected the vendors with whom we held conversations (in French or, when initiated by the vendor in English) to represent as much as possible the variety of trade that was on offer. In towns that were re-surveyed, we ensured that individual traders were not included twice.
Fig. 1. Map of Morocco showing the towns that were surveyed between April 2013 and December 2014 for the presence of reptiles used for medicinal purposes.
While individual vendors were not necessarily able or willing to provide information on all species on display, unlike the report of Fogg (1941), we communicated with a large number of vendors and herbalists who allowed us to build up a narrative of the trade and the uses of reptiles. We made most species identifications in the market itself. However, over the course of the study, we were able to photograph a large number of items, thus allowing us to verify our market identifications. We visited the markets of Marrakesh, Casablanca, Meknes, and Fez multiple times, providing the greatest amount of data on the use of reptiles for medicinal purposes. We use data from these four towns to provide a more detailed account, including more detailed analysis, of this trade.
Data on the import of African rock python (Python sebae Gmelin, 1788) and Nile crocodile (Crocodylus niloticus Laurenti, 1768), two species that are not, or are no longer, native to Morocco, were obtained from the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) trade database (http://www.cites.trade.org). These two species are listed on the CITES appendices and all import and export between parties needs to be reported to the CITES Secretariat annually. Morocco is a party to CITES, as are all countries in the ranges of the African rock python and Nile crocodile. We retrieved both data on the import as reported by Morocco and on the export to Morocco as reported by other countries.
Several species that are sold for medicinal purposes are also traded for other uses (e.g., to be used as pets) and it was not always possible to establish their intended purpose with certainty. In the few cases were there could have been uncertainty about usages, or when animals were offered for sale for potentially multiple purposes, we made an educated guess using contextual data to include or excluded these observations to eliminate questionable data. Thus, a live chameleon for sale in a shop selling live tortoises is most likely to be sold as a pet, but a live chameleon for sale at the store of an herbalist is more likely to be used for medicinal purposes. Likewise, tortoise carapaces on display at a medicinal shop suggest use in medicine, whereas the same carapace on display in a tourist shop is likely to be sold for decorative purposes. In the market, we requested the asking prices for the most commonly traded species and converted these to US dollars (US$) using an exchange rate of nine Dirham (MAD) to the dollar. We used these prices to calculate the value of each species in trade, and summed them to arrive at an overall value of the trade. Asking prices will go down after bartering or when more than one specimen at a time is purchased. Because we did not want to make any purchases, the overall figure we present in our results may overestimate the total value of the trade.
For each species in each market, we calculated the average number of species that we observed per survey (i.e., dividing the total number of individuals observed in a particular town by the number of surveys that were conducted in that town). We used four cities that were surveyed multiple times (i.e., Casablanca, Fez, Marrakesh, and Meknes), which we believe are representative of different aspects of the trade in reptiles in Morocco, to compare markets in more detail. For each market city and for each survey, we calculated the total number of species for sale as a measure of species richness, as well as the proportion of trade in live animals versus dried animals; differences between market cities were tested using ANOVAs.
We calculated minimum turnover of Mediterranean chameleon and Bell’s Dabb lizard two commonly observed species, by repeat visits to eight individual shops for the chameleons (four shops in Marrakesh, two shops in Fez and Rabat) and six individual shops for Bell’s Dabb lizard (four shops in Meknes, two shops in Marrakesh), all within a four-week period. It was not possible to calculate turnover when in between surveys a new consignment of animals had arrived and the number of Mediterranean chameleons or Bell’s Dabb lizards in the shops had increased; our turnover figures thus are minimum estimates. Using the four-week turnover figures, we calculated annual numbers of turnover for each city in which we observed Mediterranean chameleons or Bell’s Dabb lizards for sale and combined them for an overall annual figure. For the other species, we were not able to collect firm data on turnover but, given the non-perishable nature of many items, we expect that at least parts of the inventory can remain unsold for several months. To reduce the effect of non-independence of data in exploring temporal patterns, we used data from surveys done at least three months apart.
For testing whether observed frequencies of occurrence were homogeneously distributed over all classes, and whether significant differences existed between the different classes, we used sequential chi-square tests. Expected frequencies were generated based on a random distribution proportional to sampling effort per market or per survey. Pearson product-moment correlation coefficients were calculated to explore correlates with trade volumes. Statistics were run in R (R Development Core Team, 2015); we present means and ranges and accepted significance when P < 0.05.