Optimality theory predicts the maximization of productivity in social insect colonies, but many inactive workers are found in ant colonies. Indeed, the low short-term productivity of ant colonies is often the consequence of high variation among workers in the threshold to respond to task-related stimuli. Why is such an inefficient strategy among colonies maintained by natural selection? Here, we show that inactive workers are necessary for the long-term sustainability of a colony. Our simulation shows that colonies with variable thresholds persist longer than those with invariable thresholds because inactive workers perform the critical function of replacing active workers when they become fatigued. Evidence of the replacement of active workers by inactive workers has been found in ant colonies. Thus, the presence of inactive workers increases the long-term persistence of the colony at the expense of decreasing short-term productivity. Inactive workers may represent a bet-hedging strategy in response to environmental stochasticity.
Lazy workers are necessary for long-term sustainability in insect societies
by Eisuke Hasegawa, Yasunori Ishii, Koichiro Tada, Kazuya Kobayashi & Jin Yoshimura
Scientific Reports 6, Article number: 20846 (2016)
http://www.nature.com/articles/srep20846
Discussion
The proposed advantage of among-worker variation in threshold of response to task-related stimuli depends on the existence of critical tasks that cannot be halted at any moment. The nests of social insects are always threatened by bacterial and fungal infections. In termites, eggs are susceptible to bacterial infections and die quickly if attending workers are removed. Termite workers lick eggs and place anti-bacterial substances on the eggs. In ants, egg licking by workers has also been observed. Non-intermittent egg care appears to be a prerequisite for the persistence of a colony among soil-dwelling social insects.
Task replacement is also common among ant workers. When workers from a colony have been removed, tasks are performed by other physical or behavioral castes. We examined the activities of the least active 10% of workers when the most active 10% of workers were active and when they were inactive. We found that the least active 10% of workers exhibited more activity when the most active 10% of workers were inactive than when they were active. This phenomenon can be explained by task replacement by inactive workers. Inactive workers perform few tasks because they have high thresholds for responding to task-related stimuli. These workers have no fatigue and are capable of performing several tasks at any time. In fact, a previous study has shown that a portion of inactive workers began to work when all of the active workers were removed from a colony. In a rare disastrous event, such replacement provides an effective means for the continual processing of critical tasks, resulting in the long-term persistence of a colony.
Task replacement is also found in the empirical data as the negative correlation between active and inactive workers. We should note that this correlation is rather weak, because we have to use partial correlation to eliminate the effects of variable colony activity levels. Other factors may also weaken this correlation. However, it is important to note that all eight colonies exhibit negative correlation, suggesting that inactive workers tend to work when active workers are resting.
There are two major hypotheses explaining task replacement: foraging for work and variable response thresholds. As evidenced by the existence of inactive workers from empirical studies, ant workers should have variable response thresholds. Variation in response thresholds has been shown to occur among several social insect species. A previous study has suggested that M. kotokui also exhibits variation in response thresholds. Therefore, the inactive workers observed in this species are likely to originate from this variability. In contrast, the foraging-for-work hypothesis predicts that task replacement should enhance the overall performance of the colony in conducting crucial tasks principally because task replacement reduces the probability of colony extinction. In the current model, the threshold is treated as an absolute value. However, recent evidence suggests that the threshold is more akin to a probability of acting. We believe that the qualitative outcomes are similar when the action is probabilistic. In the future, we plan to test the outcomes of the model by using probabilistic thresholds as opposed to absolute thresholds.
Division of labor and age polyethism are exhibited by several social insect species. These conditions may have a profound effect on the predictions of the model. However, the current model has two states (performing a task or not). Therefore, we could not compare the outcomes of the model with the empirical data even if data on polyethism were available in the empirical data. This perspective would be an interesting extension of the current model.
In this study, both the model and empirical data indicate that the variable threshold system permitted the continuous processing of crucial tasks (e.g., egg care). Several proximate mechanisms have been proposed for the variable threshold system. We propose that the persistence of a colony is an ultimate (evolutionary) benefit for the maintenance of variable thresholds. Thus, the persistence of a colony is the critical measure of colony fitness. This principle is functionally equivalent to the concept of geometric mean fitness used to measure reproductive success under stochastic environments. Geometric mean fitness is the maximization of long-term growth rate that reduces the probability of extinction. Similarly, the persistence of a colony is equivalent to avoiding colony extinction. However, the causal factors of the two fitness measures are different. Under the concept of geometric mean fitness, environmental or demographic stochasticity increases the probability of extinction. These factors are external to the functional unit of adaptation, namely, individuals. In contrast, under our concept of colony persistence, the failure to perform critical tasks continuously is the factor that increases the probability of colony extinction. Thus, in contrast to the concept of geometric mean fitness, the factor increasing extinction probability is internal and the functional unit of adaptation is the colony. Additional studies are needed to examine “apparently suboptimal” phenomena (e.g., the extensive widespread occurrence of inactive workers) from the perspective of long-term persistence.
Our results demonstrate that ant colonies adopt a bet-hedging strategy against the unavoidable demographic stochasticity of task processing so that the colony can persist in the long run. Another common form of bet-hedging is the trade-off between growth and reproduction (e.g., the trade-off between colony growth and alate production). Our current findings differ from these studies in two major ways. First, our study addresses worker adaptation from the perspective of long-term colony survival, irrespective of the trade-off between colony growth and reproduction. Second, current adaptation is opposed to demographic stochasticity within a colony and to environmental stochasticity. Bet hedging as an adaptation is not the sole explanation for inactive workers. Indeed, there are several other non-mutually exclusive explanations for the existence of inactive workers.
In typical colonies, there should be a great deal of variation in the total workload of a colony. Previous studies have shown that the removal of some active workers increases the activity level of the remaining active workers but does not affect the inactive workers. Only when all of the active workers are removed will the remaining inactive workers begin to work. These results indicate that inactive workers are not present to serve as daily or seasonal workload replacements but instead provide insurance against a catastrophic disaster in the event that all of the active workers are not able to engage in crucial tasks. In such a scenario, these active workers are likely to be incapable of working, owing to tremendous fatigue. Thus, our findings suggest that inactive workers may primarily function to safeguard against the risk of colony extinction that may occur very rarely (e.g., once in the lifetime of a colony if it were ever to occur).
Our results may have implications for the organization of societies in general. Specifically, a society without any reserves will be unable to persist over the long term. Natural selection may only “satisfice” short-term productivity instead of maximizing it. Therefore, all existing long-lasting societies should be adapted to long-term sustainability using bet hedging or preparing reservoirs of lazy workers.
「働かないアリ、集団の絶滅防ぐ」…北大発表
読売新聞
http://www.yomiuri.co.jp/science/20160216-OYT1T50149.html
アリの集団が長期間存続するためには、働かないアリが一定の割合で存在する必要があるとの研究成果を、北海道大の長谷川英祐准教授らのチームが１６日、英科学誌サイエンティフィック・リポーツに発表した。
長谷川准教授は、「普段働かないアリがいざという時に働いて、集団の絶滅を防いでいる」と話す。
これまでの研究で、アリの集団には常に２～３割、ほとんど働かないアリが存在することがわかっている。働くアリだけを集めても一部が働かなくなり、働かないアリだけを集めると一部が働き始めるが、その理由はナゾだった。
チームは、様々な働き方のアリの集団をコンピューターで模擬的に作成、どの集団が長く存続するかを調べた。その結果、働き方が均一な集団よりも、バラバラの集団の方が長く存続した。働くアリが疲れて動けなくなった時に、普段は働かないアリが代わりに働き始めるためだ。
実際に８集団１２００匹のアリを観察すると、働くアリが休んだ時、それまで働いていなかったアリが活動し始めることが確認できたという。
(sk)
読売新聞の記事は本質をついていないような気がするけれど、新聞の記事はこれでいいのかもしれない。
いずれにしても、とても挑戦的な論文で、感じがいい。