An average of 12.2% of An. gambiae s.s. resting indoors were P. falciparum CSP positive from southern Mali in October 2009. Both molecular forms were CSP ELISA positive and there was no differential infection rate among molecular forms. Significantly more Mopti chromosomal forms (30.4%) were positive than were the Savanna (15.1%) and Bamako (9.2%) chromosomal forms. As the Mopti chromosomal form corresponds to M molecular form in Mali in most cases , finding that both M molecular and Mopti chromosomal forms were significantly associated with P. falciparum infection is not surprising. Site-specific differences in the number of CSP positive chromosomal form infection between Kela and Sidarebougou were also observed. In particular, the most common chromosomal form in each village, Bamako in Kela and Savanna in Sidarebougou, was least likely to be CSP positive.
The insignificant infection prevalence in the M molecular form in southern Mali corroborates with other studies from Cameroon and Senegal that reported no differences in P. falciparum infection between M and S molecular forms [6, 46]. A recent P. falciparum susceptibility assay among An. gambiae s.s. molecular forms from Senegal found significantly higher numbers of P. falciparum oocysts and sporozoites in the S molecular form than in the M form . This study analysed field-collected specimens of an unknown-age structure that were naturally infected with P. falciparum, whereas the Senegal study  collected eggs from the field and allowed the surviving adults to feed directly on a membrane with P. falciparum to standardize age and potential for infection. These conflicting findings may result from the origin of field collected samples (Mali vs. Senegal), to different techniques (natural vs. artificial infection), or to a differences in the age structure of the samples. Field studies from multiple sites and over multiple sampling periods are necessary to confirm the observed patterns.
Cryptic genetic differences in An. gambiae s.s. among sample sites can also limit comparisons among the previous studies [6, 46] and the present study. Genetic subdivisions beyond the M and S form designations have been reported and it is possible that differences among these subdivisions include genes associated with differential response to parasite infection. For example, recent studies in Cameroon demonstrated a subdivision in the M molecular form into discrete Forest-M, characterized as M molecular form and Forest chromosomal form (fixed for standard gene arrangement) and Mopti-M populations with typical Mopti karyotypes [12, 21]. In analyses using SNPs from immune signaling genes, three genetically distinct An. gambiae s.s. populations were observed in Mali: the M molecular form, the S molecular form (S1), and a subdivided S Pimperena form (S2) . Further, SNPs associated with P. falciparum infection were differentially distributed among M, S1, and S2 populations . Of interest, data presented here is similar to that reported in Riehle et al where a cryptic subgroup of An. gambiae, indistinguishable in molecular form but distinguishable via microsatellites amplified from chromosome 3, was susceptible to P. falciparum.
Differences in the local environment may likewise affect associations between An. gambiae forms and P. falciparum infection. For example, Dolo et al demonstrated that irrigated zones of Mali allowed for constant CSP positivity across seasons along with low human blood feeding and sporozoites indices, whereas in the non-irrigated zones, CSP positivity fluctuated seasonally, being high in the wet season and low in the dry season. Dolo et al hypothesized that malaria prevalence in villages adjacent to irrigated rice fields is consistently low in this environment because adult density is inversely related to blood feeding due to high mosquito densities driving villagers to protect themselves with repellants and bed nets [49, 50]. The Bamako chromosomal and S molecular forms were dominant in Kela (~3 km to a river), whereas the Savanna chromosomal form and S molecular form predominated in Sidarebougou (~10 km to agriculture fields). Kela is located close to a river that has the ability to flood and create additional oviposition sites not in the dry season likely increasing mosquito densities (11.2% prevalence), whereas in Sidarebougou mosquitoes densities are likely dependent on the wet season (13.9% prevalence). These habitat differences could contribute to the genetic variation (and potentially phenotypic variation) observed at different locations and, as Dolo et al hypothesized, habitat may play a role in the vector ecology of An. gambiae. Collectively, these studies highlight the importance of both genetic and environmental determinants of susceptibility to infection.
There were statistically significant P. falciparum differential infection rates among chromosomal forms and trends among chromosome inversions. Mopti had the highest CSP positivity (30.4%), followed by Savanna (15.1%) and Bamako (9.2%) forms. The data presented here indicated that Mopti chromosomal forms (2Rbc/u) were more likely to be positive for P. falciparum CSP. A previous study in Kenya identified a significant association between the total number of inversions and a decreased likelihood for P. falciparum infection . Data presented here did not show this specific association to be statistically significant, but standard chromosomal arrangements tended to increase the likelihood of being CSP positive. In particular, in Kela where the Bamako form was dominant and least likely to be positive, mosquitoes with standard or heterozygous arrangements were more likely to be CSP positive than mosquitoes homozygous for 2Rjcu and 2Rjbcu (Bamako form) (Table 3).
A number of studies have examined associations of chromosomal polymorphisms with malaria infection. A small region of chromosome 2 L has been associated with infection susceptibility, regardless of P. falciparum genotype . Genes within this region encode for melanization or parasite encapsulation . Within the 2La region, the APL1 gene, which encodes for natural resistance to P. falciparum, exhibited extremely low genetic diversity within the M molecular form, but high diversity in the S molecular form that may have arisen from larval infection [26, 35]. Alternatively, higher diversity at the APL1 locus in the S molecular form may be associated with a more diverse array of responses to P. falciparum and reduced susceptibility to infection. Similar results were identified within the anti-parasite and anti-bacterial gene TEP1r. Specifically, TEP1r was diverged between the M and S molecular forms and one variant showed a strong association with resistance to infection with a rodent malaria parasite and with P. falciparum. Additional studies comparing An. gambiae molecular and chromosomal forms with P. falciparum infection that incorporate An. gambiae speciation, genetic diversity at immune loci (e.g., TEP1r, APL1, Toll5B) as well as larger temporal and spatial scales may help to extend findings reported in this study.