是害蟲也是獵物-探究鼠類在避敵反應和農業生態系中的雙重角色

Of pests and prey: investigating the dual role of rodents from their responses to predation risk to their involvement as pests in agriculture

王意安(Ian Nicholas Best)

指導教授 ： 郭奇芊 ； 裴家騏

國立臺灣師範大學/理學院/生命科學系/博士(2021年)

https://doi.org/10.6345/NTNU202100227

摘要

鼠類，一個無所不在又分布廣泛的分類群，是許多捕食者的獵物，因此有許多研究試圖研究鼠類被掠食的風險。這些小型哺乳類動物具有躲避掠食者的行為，包括辨別掠食者氣味以避免被取食。然而，不同物種對風險的反應不同，反應機制可能與生理、形態以及生活史特徵有關。在不同的狀況下，物種的行為可能與生活史步調(Pace-of-life)有關，生活史步調從快到慢，以及行為由大膽到謹慎。另一方面，因為齧齒類會造成全球農業巨大損失，也有研究試圖結合被獵食風險來降低鼠害；相較於傳統上利用的化學防治方法，如滅鼠藥的利用，這種方法對環境較為友善，也可降低對非防治目標野生動物的傷害。
本論文的一大部分在檢測掠食者氣味，引發野外族群不同鼠種躲避掠食者行為。另一部分則是調查一個農業地區鼠類對不同作物的影響，以及農民利用滅鼠藥和其他化學防治的比例。
第一個研究計畫中(第二章)，我探討在台灣東部的花蓮，四種鼠類暴露在不熟悉環境，以及非共域掠食者—石虎(Prionailurus bengalensis)氣味後的行為反應。這四種老鼠包括三種原生鼠種(田鼷鼠Mus caroli, 赤背條鼠Apodemus agrarius, 小黃腹鼠Rattus losea)，和一種外來入侵種(緬甸小鼠Rattus exulans)。這些老鼠被放置在實驗室內進行連續兩晚的實驗。結果發現鼠類面對掠食者氣味，避敵行為的時間並不會增加。然而，面臨風險，物種間的行為有所差異，體型較小的物種較為大膽，較大的物種則較為謹慎。結果符合生活史步調假說，生活史特徵和抵抗掠食者行為有關。
第二個研究計畫，我使用放棄密度(giving-up density)實驗，配合自動照相機的使用，在有石虎出沒的苗栗縣，研究間接(植被覆蓋程度)與直接(掠食者氣味)被捕食風險訊息如何影響野外鼠類的覓食行為。比較包括原生石虎、引入的家貓(Felis catus)和台灣沒出現的短尾貓(Lynx rufus)這些掠食者的氣味對於老鼠群聚是否有不同的影響。結果顯示，老鼠造訪食物站和取食的次數，以及種子被取食程度，不會受到任何一種掠食者氣味的影響，但卻會受到微棲地所影響：和空曠暴露的地方相較，有植被覆蓋的棲地，老鼠取食較多種子。另外，自動照相機發現，體型小的鼠種(A. agrarius)在行為上較體型大的物種(R. losea)大膽。本章結果和第一章結果類似，同樣較支持行為與生活史步調有關。
在第四章中，我訪問苗栗縣農業地區的農民，鼠類危害的程度，相關農害防治措施，農民對鼠類的態度，以及改變農害防治方法的意願。結果顯示老鼠對於稻米的危害最為嚴重，對蔬菜和水果則較無害。此外，只有三分之一的農民指出他們目前有使用滅鼠藥，但有三分之二的農民有使用其他種農藥。是否使用滅鼠藥和種植作物種類以及鼠害程度有關。和預期相符，當農民覺得鼠類危害很大時就比較會使用滅鼠藥。儘管使用滅鼠藥的農夫不多，但是大部分受訪者對於老鼠持負面的觀感，且與鼠類的危害程度相關。此外，對鼠類有負面觀感的農夫更傾向於會使用滅鼠藥，因此，過去鼠害的經驗與農人的態度決定了防治的措施。同時，滅鼠藥的使用可能反應了鼠害相當嚴重。另一方面，農民雖然支持減低滅鼠藥和殺蟲劑的使用，但並不願意完全不使用這些化學藥劑。適當的獎勵措施可能使農民採用生態友善農法。
本論文的第二章及第三章發現，掠食者的氣味，即使是原生石虎的氣味，也不會引發老鼠的禦敵行為和抑制覓食活動。種間，甚至是個體間行為的差異，會影響對風險的反應。因此，利用創造「恐懼地景」來作為生態防治策略可能不是最有效的方法，未來還需要更加了解如何利用掠食者風險為基礎的鼠類防治策略。同時，不同鼠種之間的行為差異，可能造成不同的鼠害問題。根據本論文第四章對農民的調查，開發整合性生態防治措施是有潛力的，但是，如何發展有效的策略還需要更多的研究。

關鍵字

老鼠 ； 石虎 ； 捕食風險 ； 行為 ； 取食 ； 大膽 ； 生活史步調 ； 害蟲 ； 態度 ； 滅鼠藥 ； 管理

並列摘要

Rodents, a ubiquitously distributed and pervasive taxonomic group, are common prey to a suite of predators, and as a result have been the subjects of countless studies manipulating predation risk. These small mammals are equipped with antipredator defenses, which includes the ability to discriminate predator odors through olfaction and subsequently modify their behavior to avoid potentially fatal encounters. How different species respond to risk may also vary between species, and these mechanisms may be linked to physiological, morphological and life history traits. Behaviors of species across different contexts may be associated with pace-of-life (POL) following a fast-slow continuum and boldness-shyness axis. Since rodents also cause devastating losses to agricultural industries on a global scale, innovative methods incorporating perceived predation risk to discourage rodents as pests have been the subject of consideration. This method could be an ecologically friendly alternative to conventional chemical approaches, such as rodenticide application, and reduce risks for non-target wildlife.
A major component of this thesis was to test the efficacy of predator odors at instigating antipredator behavioral responses in different species of rodents from wild populations. Another component of this thesis was to survey an agricultural area to understand the impact rodents have on various crops and what proportion of farmers use rodenticides and other chemical-based pest management.
For my first project (Chapter 2), I investigated the behavioral responses of four rodent species in Hualien County, of eastern Taiwan, exposed to a novel microenvironment and cue from an allopatric predator, the leopard cat (Prionailurus bengalensis). Three of the wild-caught rodent species were native (Mus caroli, Apodemus agrarius, Rattus losea) and one invasive (R. exulans) and all were subjected to two consecutive nights of experimental trials in a laboratory setting. Rodents did not respond to the predator odor by increasing the amount of time exhibiting defensive behaviors. Instead, inter-specific behavioral variation was observed exemplified by the smaller species performing more behaviors for boldness and the larger species exercising more caution in response to the risky contexts. These results are in accordance with a growing consensus that behavior is linked to pace-of-life (POL); differences in life history traits are associated with behavioral traits following a fast-slow continuum.
For my second study (Chapter 3), I conducted a giving-up density (GUD) experiment complemented with camera traps in Miaoli County, where leopard cats are found, to investigate how indirect (e.g. vegetation cover) and direct (e.g. predator odors) risk cues influence the foraging behavior of wild populations of rodents. I tested whether the odors of the native leopard cat, introduced domestic cat (Felis catus), and exotic bobcat (Lynx rufus) would have differential effects on rodent communities. Visitation, foraging and seed consumption at the experimental food patches were not affected by any of the predator odors, but instead by microhabitat type; rodents consumed more seeds in food stations under vegetation cover compared to exposed stations. Additionally, based on the camera trap data, the smaller species (A. agrarius) demonstrated a higher proportion of behaviors conferring boldness compared to the larger rat species (R. losea). These results are similar to those from chapter one, and also largely adhere the hypothesis that behavior is linked to POL.
In Chapter 4, I describe my study where I conducted a survey for farmers in an agricultural area of Miaoli County to investigate the extent of pest-related damage caused by rodents, pest control practices employed by farmers, their attitudes toward rodents, and willingness to change their pest management. Rodents were reported to be the worst pest for the crop category rice, but not vegetables nor fruit. Furthermore, only about one-third of farmers indicated they currently use rodenticides, whereas around two-thirds apply other types of pesticides. The most important predictors for rodenticide use were type of crop grown and extent of problems caused by rodents. Intuitively, farmers were more likely to use rodenticides if they perceived rodents to be very problematic and cause damage. Despite the low frequency of reported rodenticide use, most participants had negative attitudes toward rodents, with extent of rodent problem having the strongest influence on attitudes. Moreover, participants with negative attitudes toward rodents were found to be more likely to use rodenticides. Therefore, experiential factors, including perceived damage caused by rodents and farmers’ attitudes toward the pest likely facilitate their pest control practices. Moreover, the use of rodenticides may be a reactive measure, indicated by the severity of problems and damage inflicted by rodents. Participants expressed support to reduce their usage of rodenticides and pesticides, but were not willing to stop using the chemical products altogether. With adequate incentives farmers may be agreeable to adopting ecologically-based pest management strategies.
In both studies described in Chapter 2 and 3, predator odors did not elicit defensive behaviors and suppress foraging activity, even in response to the native leopard cat cues. Behavioral variation at the species level, and even between individuals, is an important factor that can influence overall responses to risk. Therefore, the application of these direct risk cues simulating the ‘landscape of fear’ as an ecologically-based management strategy for rodents as pests may not be the most effective and will require further attention to better understand the capabilities of predation risk-related pest control. Concomitantly, inter-specific behavioral variation occurs between different rodent species, therefore, their roles as pests may not all be equal. Based on the farmer’s responses from the survey discussed in Chapter 4 there is potential for the integration of ecologically-based management practices, however, developing effective strategies will require more attention.

並列關鍵字

rodents ； leopard cat ； predation risk ； behavior ； foraging ； boldness ； fast-slow continuum ； pace-of-life (POL) ； pest ； attitudes ； rodenticide ； management

參考文獻

Adler, G. H. (1995). Habitat relations within lowland grassland rodent communities in Taiwan. J. Zool., 237(4), 563-576.

Akaike, H. (1974). A new look at the statistical model identification. IEEE T. Automat. Contr., 19(6), 716-723.

Albert, A., & Anderson, J. A. (1984). On the existence of maximum likelihood estimates in logistic regression models. Biometrika, 71(1), 1-10.

Alexander, R. D. (1974). The evolution of social behavior. Annu. Rev. Ecol. Evol. Syst., 5(1), 325-383.

Ann, P. J., Lee, H. L., & Huang, T. C. (1999). Brown root rot of 10 species of fruit trees caused by Phellinus noxius in Taiwan. Plant Dis., 83(8), 746-750.