Accurate identification of mosquito species is necessary for determining the composition of vector populations, particularly as this changes in the face of differential selective pressure exerted by vector control measures, such as insecticide-treated nets [1, 2] or indoor residual sprays . These vector control measures occur mainly where morphologically indistinguishable species co-exist as vector complexes, such as Anopheles gambiae s. l., which dominates malaria transmission in most of Africa. Estimating mosquito age distribution of mosquito populations is also crucial for assessing their capacity to transmit malaria and other pathogens [4, 5]. For example, a population dominated by young mosquitoes indicates a successful vector control with ITNs and IRS interventions which reduce longevity and therefore both density and infectiousness [6, 7]. Only anophelines that are at least eleven days old can transmit malaria parasites due to the period required by the parasites to develop inside the mosquito , so even modest reductions of mean survival rates within vector populations can deliver substantive epidemiological impact [4, 9–11]. Assessing mosquito population age structure prior and subsequent to control interventions therefore provides a strong indication of the efficacy of that intervention.
Several techniques have been established to estimate the age of Anopheles mosquitoes. These include the traditional techniques that involve observation of the morphological changes that occur in the reproductive system of the female mosquitoes to estimate their physiological age [12–16]. Recently biochemical approaches based on age-related changes to the abundance of cuticular hydrocarbons  and gene transcription [18–21] have shown considerable promise although they may also be costly  and, therefore, have limited application for large-scale ecological or epidemiological studies. Additionally, sibling species identification within the critically important An. gambiae complex and the Anopheles funestus group from Africa relies almost exclusively upon standard Polymerase Chain Reaction (PCR)  and multiplex PCR [24, 25] protocols. These protocols are nonetheless time consuming and can be costly in a resource limited area. It is for this reason that in most cases, only a small sample of the population is tested to estimate species distribution in an area. More recently, near-infrared spectroscopy (NIRS) has been developed as a complimentary age grading and species identification tool for Africa's main malaria vectors An. gambiae s.s. and Anopheles arabiensis mosquitoes [26, 27]. NIRS is a rapid, non-destructive tool that can determine age and species of hundreds of mosquitoes per day. No reagents are required and only basic computer skills are needed. This NIRS technique is more cost-effective than PCR after about 7,000 samples have been analysed . However, for mosquitoes, this tool has only been applied to fresh anesthetized samples, limiting its use particularly in large-scale studies where preserving samples collected under widely-dispersed sampling sites is often essential.
Current methods to preserve mosquitoes for DNA extraction or dissection include desiccation and stabilization in various storage buffers. Preservation by desiccation involves complete dehydration of samples over silica gel beads and cotton wool. Silica gel must be kept activated but is widely relied upon particularly in large-scale studies in tropical field settings for preserving mosquitoes prior to DNA and antigen assessment by PCR and ELISA techniques. For subsequent analysis of samples with NIRS however, it is also key that samples are preserved in a way that minimizes chemical degradation. Specifically, NIRS is thought to differentiate sibling species of An. gambiae s.l. (An. gambiae s.s. and An. arabiensis) based on the composition of cuticular hydrocarbons, but may also rely on water content which is known to differ between these two sibling species . Additionally, age-grading of these species depends on gene transcripts [20, 21] and change of a range of other bio-molecules including cuticular hydrocarbons . While desiccants are a low-cost alternative to preserve insects for DNA analysis, desiccants must be kept activated and the suitability of insects for dissecting can be poor .
Insects are also commonly preserved by suspending in solvents such as ethanol, but studies of this approach for samples to be assessed with NIRS indicate a slight reduction in accuracy relative to scanning fresh samples . Also, solvents leave samples brittle and thus they are difficult to dissect. Other storage procedures used to store biological samples include ultra-cold storage in liquid nitrogen but the costs required for maintenance of liquid nitrogen is prohibitive in most field settings, particularly in resource limited tropical countries.
® (Ambion, Inc., Austin, TX) is an aqueous, non-toxic storage reagent that has been used to preserve mosquito DNA  and other samples at room temperature up to one month , and indefinite storage time is possible if held at -20°C. Although RNAlater
® is more costly than solvents or desiccants, samples are suitable for DNA extraction and dissection. However, NIRS has not been used to analyse mosquitoes stored in RNAlater
®. Since RNAlater
® is currently being used by some researchers to preserve mosquitoes and has some advantages over desiccants and solvents for subsequent DNA analysis and dissection, the objective of this study was to compare the accuracy of NIRS for determining the age and species of freshly anesthetized An. gambiae s.s. and An. arabiensis to those preserved in RNAlater