The effect of temperature on scouting turnover in this study aligns with the prior studies that found temperature to influence ants foraging behaviour (Crist and Williams, 1999; Ruano, Tinaut and Soler, 2000; Van Oudenhove et al., 2011; Bishop et al., 2014). Frank and Linsenmair (2017) confirmed this is also true with M. analis. Since M. analis rely on trail pheromone, which decays rapidly with the temperature rise (Bayliss and Fielding, 2002; Yusuf et al., 2014), the temperature is arguably among the driving factors of M. analis foraging. Apart from trail decomposition, ants, in general, are susceptible to desiccation due to their higher surface-to-volume ratio (Kumar and O’Donnell, 2009). Megaponera analis is relatively large (Villet, 1990) and is characterised by its melanism in its exoskeleton (Clusella Trullas, van Wyk and Spotila, 2007) and, therefore, prone to overheating by radiation (Kaspari et al., 2015) due to quicker absorption of solar radiation (Law et al., 2020). On average, M. analis has been reported to prefer a temperature of between 23.10 - 25.40 °C when observed in Kenya, Cameroon and Tanzania (Patel et al., 2018). The preferred temperature range of 23.10 - 25.40 °C could be the reason the distribution of these ants in Africa is within the tropical regions. In Limpopo Province, where this study has taken place, sunrise is as early as 05h00, and temperatures can reach 25 °C within a few hours from sunrise and set at 19h00 during summer time (Mulaudzi, Maluta and Sankaran, 2015). This preferred temperature will mostly be experienced after sunset except for cloudy days. Given this, it is likely the case that bimodal scouting in this species evolved as a measure to avert rising temperatures. This will not be the first species to do this, in workers of Myrmecia croslandi, foraging patterns changed from continuous to bimodal when there were temperature changes (Jayatilaka et al., 2011). A sister species to M. analis, Dinoponera gigantea, also has a bimodal foraging pattern, preferring mornings and late afternoons into the night (Oliveira et al., 2002). Ophthalmopone berthoudi, which shares a clade with M. analis (Schmidt, 2013), also has bimodal foraging activity (Crewe and Peeters, 1987). However, Dinoponera gigantea is found in the Brazilian rainforest (Oliveira et al., 2002), and the temperature regime at the base of the forest is known to be substantially cooler and stable as compared to areas without canopy, which are characterized by mercurial temperature (Kosheleff and Anderson, 2009). Despite the stable forest ground temperature, D. gigantea workers limit foraging during the mid-day hours to mornings and late afternoons (Oliveira et al., 2002). Given the temperature stability on the forest ground, if the temperature alone influenced foraging, epigeal ants found in tropical forests like D. gigantea would not have bimodal foraging. The mean temperatures in this study showed that the evening had higher temperatures than the night (Figure 3B), but despite this, the evening had more scouting trips than the night (Figure 4). This suggests that temperature alone is not the key factor of scouting turnover. Humidity was found to have a direct proportional relationship with the scouting turnover (Figure 3C), and the highest scouting trips recorded in the morning were accompanied by the highest mean humidity in the mornings (Figure 3D). In Tanzanian populations of M. analis, a study found rainfall and subsequently humidity to be the driving factor in M. analis foraging. These results support our result that humidity also play a role in influencing scouting turnover. Given the aforemnetioned discussion about the temperature and scouting turnover, we suggest that humidity and temperature are working in synergy to influence scouting turnover. Humidity is important as it help both the ants and termites from desiccation but termites benefit more by not having to provide the entire moisture for gallery construction from its gut (Shanbhag and Sundararaj, 2012). On the other hand, it is also beneficial energy-wise for M. analis to find freshly built galleries as this will save them the energy needed to cut the galleries open. Despite humidity having a positive correlation with scouting trips and temperature having a negative correlation, the evening had both the higher mean temperature and lower mean humidity as compared to the night (Figure 3B & 3D), but still had the highest scouting trips as compared to the night, suggesting that the night is not the favourable for scouting turnover despite favourable microclimatic conditions. Our box plots shows that there is a rapid scouting turnover at the beginning of the session (Figure 2A), suggesting that hunger state is influencing the scouting turnover. After a while when scouts did not return to the nests, colonies delayed turnover with the longest turnover being 124 minutes. We suggest that this delayed turnover is used as a tradeoff in case all scouts had experienced mass predation. The poor number of scouting trips even under the favourable microclimatic conditions suggests that the bimodality of M. analis is most probably driven by hunger status instead of microclimatic conditions alone. Bayliss and Fielding (2002) observed outliar scouting activity of this species which happed during midday while Longhurst and Howse, 1979a observed scouting commencing from 15:00 local time. This observation shows that when the colony is starved, M. analis will avert the bimodality to secure food and this support our suggestion that scouting is influenced by the hunger status. We believe the high scouting activity and subsequent raids in the evening supply enough food to last the colony for the entire night until such food is completely metabolised by the morning, hence another sharp peak in scouting turnover in the morning. Our analogy is that the morning peak in scouting turnover is intended to have multiple raids in the morning so that there could be enough energy until the evening since day temperatures are unfavourable for both the ants and the termite prey and that the evening scouting turnover peak is intended to have multiple raids in the evening to get enough energy until the morning. Indeed, food searching does not always produce the desired results, and this is likey the reason there are sporadic scouting trips during the night and afternoon. The results of scouting turnover per reserve across various observation periods support the latter argument by having all afternoon observations with the quickest scouting turnover (Figure 2C). Since bimodal behaviour is also practised by Ophthalmopone berthoudi (Crewe and Peeters, 1987), which is the only species sharing a monophyletic clade with M. analis (Schmidt, 2013), we can safely say this behaviour is an inherited behaviour from the common ancestor. This trait has probably evolved to match the adaptation and foraging strategy of some termites which are sensitive to heat. Generally, termites are sensitive to high temperatures like most insects but in the study of four termites species from different genera, including the genera Odontotermes which is among the preffered prey by M. analis (Longhurst, Johnson and Wood, 1978), exposure to sunlight increased mortality of termites (Hewitt, Nel and Schoeman, 1972). This imply that termites will directly avoid too much sun exposure and this behaviour was likely reciprocated by M. analis as part of adaptive strategy in order to optimize foraging. Evolutionary, M. analis would not have survived natural selection if their foraging times goes against the foraging time of their prey. In the study of another ponerine species Pachycondyla commutata (Mill, 1984), it was also found that scouting peaked between 06:00 and 10:00 as well as between 14:00 and 20:00. The number of raids substantially decreased as the night progressed. These results agree with our results that scouting trips was highest in the morning, evening and afternoon respectively with the night experiencing the lowest activity (Figure 4). Both Yusuf et al., (2014) and Mill (1984) found that the night had the lowest termite catches and should foraging continue at nights such foraging would not be optimal (Pyke, 1984). Moving scouting activity to the morning was more likely an adaptive foraging strategy to optimise foraging and collect more termites with minimal energy. Concerning the costs of raiding, literature had already established that it is energy costly for the colony to send one column out (Duncan, 1995). Mill (1984) suggests that the reason why scouting peaks during morning and evening is that the termite soldiers are preparing withdrawal back to the nest or were preparing to go out of the nest due to the nocturnality of the species from the Syntermes genera. Like most predators, M. analis is known to not store their prey for later consumption. In congruent to our results, Yusuf et al., (2014) reported that more raids (56%) were recorded in the morning as compared to the evening raids which were 42% whereas nocturnal raids were seven percent (7%). Study that looked at the predation of this species across five different termite species revealed that factors which influence raiding preference on certain species includes abundance, foraging pattern of termites prey, prey size and defence strategies employed by the termites (Longhurst, Johnson and Wood, 1978). Typically M. analis colonies release between 10 to 20 scouts but when colonies are starved, they have been observed to release between 30 to 50 scouts at a foraging session (Frank and Linsenmair, 2017b). This sharp increase in the number of scouts support our suggestion of hunger being the deciding factor of scouting turnover. A study on Temnothorax rugatulus revealed that confined spaces increase metabolic rate (Cao and Dornhaus, 2008) and since M. analis prefers confined nesting spaces, there is a chance the food collected by the evening is fully metabolised by the morning. The latter could be among the reasons explaining this species high appetite. To give a picture of their appetite, the study at the coastal dry forest in Tanzania found a mean predation rate of 932 termites per day by M. analis (Bayliss and Fielding, 2002). In addition, reproduction in M. analis is energy expensive due to its monogyny. Since the queen has high fecundity (Schmidt, 2013), her energy requirement is reasonably high. It was already established that ant larvae’s food demand regulates the foraging behaviour in the case of Cerapachys biroi (Ulrich et al., 2016) and M. analis colonies have larvae continuously due to the continously egg production. Given these reasons, it is understandable why the nest participates in multiple raids per foraging session. The reserves at the more arid part of the study area viz. Maremani and Musina nature reserves had the quickest turnover as compared to the reserves in the Waterberg region (Figure 2C). Maremani and Musina nature reserves are characterised by poor basal covers (Curlewis et al., 2018) as compared to the Masebe and D’nyala nature reserves which have rich grass, forbs and herbs diversity (Mundalamo, 2019). Basal cover influences the microclimatic aspects of the foraging area by creating shade, leaf litter, and suppressing wind erosion. Basal cover helps maintain soil temperature, moisture and relative humidity while its absence is associated with harsh microclimatic fluctuation including winds and rise in temperatures (Ewers and Banks-Leite, 2013). The quick turnover at these reserves could be an adaptive strategy to avoid desiccation since barren land is hot and dry. Importantly, since the poor basal cover affects termite abundance, quick turnover at these reserves is more likely due to the hunger following unsuccessful raids owing to prey scarcity. A study about the foraging behaviour of leaf-cutting ants found that changes in microclimatic conditions changed the foraging behaviour of Atta cephalotes (Bustamante, Amarillo‐Suárez and Wirth, 2020). In the study of Lasius niger, it was found that the more the scouts are starved, the quicker they get out of the nest for scouting (Mailleux et al., 2011) and this provide enough evidence for our suggestion that the bimodality and scouting turnover of M. analis is driven mainly by the hunger state apart from the microecological conditions.