台灣中部茄園蟲害管理之群聚生態學探討

Abstract

本論文從群聚的角度探討茄子(Solanum melongena L.)上植食性及捕食性昆蟲與蟎類在茄園內的群聚結構，採用科為分類的基本單位，取樣單位為葉、花、果實三部位，其中葉部又分為未展開葉及展開葉二部位。依相對豐量及群集分析的結果顯示，常見植食性種類有薊馬科(Thripidae)、蚜蟲科 (Aphididae)、粉蝨科(Aleyrodidae)、葉蟎總科(Tetranychoidea)及細蟎科(Tarsonemidae)。捕食性種類以捕植蟎科（Phytoseiidae）為主要種類，其次為花椿象科 (Anthocoridae)、盲椿象科 (Miridae)及癭蚋科 (Cecidomyiidae)。葉蟎總科及粉蝨科在中、老葉的分布較多；細蟎科主要棲息在嫩葉及果實花萼內為害，造成葉片組織厚化，阻礙茄株的生長；薊馬科則在葉、花及果實上均有很高的發生率，屬全株型的害蟲。茄花若受薊馬取食為害，會造成落花或果實表皮上有褐色條斑的食痕，若於果實期為害，茄果會呈彎曲之畸型狀或表皮上有白色條斑或褐色條斑，影響市場的商品價值甚大，因此，認為薊馬科及細蟎科應列為茄株上不可輕忽的二大重要有害生物。薊馬科主要發生種類為南黃薊馬 (Thrips palmi Karny)，細蟎科僅發現茶細蟎 (Polyphagotarsonemus latus Banks)一種。 連續四年在不同栽培管理模式的茄園內進行昆蟲蟎類群聚多樣性的研究，顯示栽培管理模式會影響昆蟲蟎類群聚的結構。從生物多樣性的分析，發現茄園除不除草管理對茄株上之物種豐富度的影響大於施藥管理者。然而，施藥次數愈多，會降低害蟲蟎類的族群數量及其均勻度，故生物多樣性指數也低。捕食性天敵在不施藥的茄園內可與其他種類的天敵維持一定的消長變化，但在施藥田的消長變化則大於不施藥區，顯示藥劑防治的確會影響茄園內捕食性天敵的活動。 茄株種植期均不施藥的情況下，以粉蟎總科(Acaroidea)及姬葉蟬科(Typhlocybidae)的生態席位寬度值最高，表示這二類害蟲在茄株各部位的分布極為均勻，而粉蝨科、潛蠅科(Agromyzidae)及葉蟎總科則最低，主要是因這三類害蟲集中分布在展開葉上，故分布最窄。細蟎科及管尾薊馬科(Phlaeothripidae)在施藥區的生態席位寬度高於不施藥區者。薊馬科、粉蝨科、姬葉蟬科、蚜科、夜蛾科(Noctuidae)、潛蠅科、葉蟎總科間的重疊性相當高，可見其在茄株上的資源利用極為相近。捕食性天敵花椿象科及盲椿象科的生態席位重疊性相當高，捕食性薊馬科僅發現印度食蟎薊馬（Scolothrips indicus Priesner）一種，其與癭蚋科的重疊度高達99.8%。現存於茄園內之捕食性天敵與細蟎科的重疊度最低，尤其在施藥區，顯示茄園內應沒有細蟎的主要天敵存在。 本論文嘗試以feeding guild的方式來整合害蟲蟎種類，將其分為咀嚼者(chewers)、刺吸者(suckers)及內吸收者(internal suckers)三大類。其中刺吸者再區分為昆蟲類(sucking-insects)及蟎類(sucking-mites)，分析其在茄株上的空間分布型，進而建立茄園蟲害管理的取樣技術。不同feeding guild的s2/m值及Lloyd’s mean crowding (m*) 隨時間變化的走勢與單一種族群極為雷同。但index of patchiness (m*/m) 分析結果則與單一族群略有不同，施藥區某些害蟲蟎類在展開葉上的指數值有小於1者，然而以feeding guild歸類者，各群的聚集度均大於1，但指數值卻較單一種族群為低，以展開葉的分析結果最為明顯。以不施藥區之密度調查資料，依Taylor’s power law分析之a與b值來估算每一取樣單位上平均密度在5-200隻的取樣數，結果顯示以展開葉所需的取樣數最多。當展開葉上密度達10隻時，估算薊馬需取257片展開葉，蚜蟲則需649片，粉蝨需209片。若以sucking-insects為單元來估算密度時，當密度為10隻下，最適取樣數為123片展開葉，低於單一種族群者。 南黃薊馬是造成茄果商品價值降低的關鍵性害蟲，故本論文也特別就溫度對南黃薊馬生活史特徵及族群介量的影響進行研究。卵在35℃下僅有4-8%可存活發育至成蟲，存活之成蟲其平均壽命僅2.8-2.9日，且無法產下任何卵粒以延續族群，由此推測35℃已接近其發育與生殖臨界高溫。南黃薊馬在15、21、25、30℃下飼養，卵至成蟲所需發育時間分別為29.9、19.6、12.3及10.4日。雌蟲壽命隨溫度的升高而縮短，分別為21.6、20.2、15.4及9.7日。雌蟲總產卵數及產卵速率以飼養在25℃下者為最大，但其產卵期最短約21日，平均每隻雌蟲的產卵數為57.1 ± 4.9粒。利用溫度與發育速率的直線迴歸模式估算發育臨界低溫為7.7 ± 0.2℃，有效積溫為227.2 ± 3.3°D。利用1999-2002年農業試驗所霧峰氣象站每日的平均氣溫估算南黃薊馬在台灣中部一年可能發生25-26代。從齡別存活率(age-specific survival rate)、齡別繁殖率(age-specific fecundity) 及齡別繁殖值 (age-specific maternity) 曲線表現來看，南黃薊馬在25℃下的繁殖速率最快。在15-25℃下，內在增殖率(the intrinsic rate of increase, r) 隨溫度的上升而增加，但在30℃時則下降，其r 值分別為0.033, 0.046, 0.157及 0.118 day-1 (15、21、25、30℃)。25℃下之平均世代時間（T）及淨繁殖率(R0)分別為18.6日及18.5 eggs，均高於15、21及30℃者。綜合以上結果顯示，25- 30℃為南黃薊馬族群增長較適合的溫度。

I studied the community structure of arthropod herbivores and predators on eggplant (Solanum melongena L.) in central Taiwan. The leaves, flowers and fruits of eggplant were sampled, among them, the leaves were divided into spire leaves and expanded leaves. The insects and mites were identified to their families and their numbers counted. Relative abundance and cluster analysis indicated that Thripidae, Aphididae, Aleyrodidae, Tetranychoidea and Tarsonemidae were common herbivores on the eggplant. Phytoseiidae was the major predator, followed by Anthocoridae, Miridae and Cecidomyiidae. Tetranychoidea and Aleyrodidae mainly inhabited on the moderate and aged leaves. Tarsonemidae preferred the spire leaves and the calyx of fruit to the expanded leaves and flowers. The color of spire leaves turned dark green and tissue became thickening after injury by Tarsonemidae, so that the eggplant growth was obstructed. Thripidae frequently occurred on leaves, flowers and fruit of eggplant and they were more prevalent than others. Injury of flowers and fruit by Thripidae caused the flowers to fall and fruits with some white or brown narrow strip scars and deformities, hence decreasing yields and marketability of the eggplant. Therefore, I suggested both of Thripidae and Tarsonemidae were two key pests on the eggplant. Thrips palmi Karny was the dominant species in Thripidae and Polyphagotarsonemus latus Banks the only species of Tarsonemidae found on eggplant. The differences in crop management practices could drastically affect the insect-mite community in eggplant fields. The family richness in the eggplant field with pesticides application was lower than the weeded one. The more the pesticides applied, the less was the evenness, and hence the biological diversity was lower. The fluctuation of predators in plot without pesticides application was more stable than that with pesticides application. The niche breadth of Acaroidea and Typhlocybidae in eggplant plot without pesticides application were the widest, indicating that these two families distributed evenly on each parts of eggplant. Aleyrodidae, Agromyzidae and Tetranychoidea mainly concentrated on the expanded leaves. Therefore, their niche breadth was narrower. The niche breadth of Tarsonemindae and Phlaeothripidae in plot with pesticides application was wider than that in the plot without pesticides application. However, those of other herbivores were in the contrary. Analysis revealed that Thripidae, Aleyrodidae, Typhlocybidae, Aphididae, Noctuidae, Agromyzidae and Tetranychoidea had higher niche overlap. This result suggested that resource utilization of these herbivores on three parts of eggplant was very similar. In predator both of Anthocoridae and Miridae had higher niche overlap. Niche overlap of Scolothrips indicus Priesner, with Cecidomyiidae was 98.8%. It is also noted that the niche overlap between all predators and Tarsonemidae was the least, suggesting that there were no predators against Tarsonemidae in the eggplant fields. From the view of injury-damage type, I tried to use feeding guild to group the herbivores as chewers, sucking-insects, sucking-mites and internal suckers in order to develop a practical sampling technique for the eggplant IPM program. The fluctuation trend of the value of s2/m and Lloyd’s mean crowding (m*) of various feeding guild weekly was the same as the ones of single population, but they were different on index of patchiness (m*/m). The index weekly was partly less than 1.0 for single population in plot with pesticides application; however, the index weekly was all more than 1.0 for feeding guild. Taking the data collected from plot without pesticides application to estimate the optimal sample size for density ranging from 5 to 200 individuals on each part of eggplant according Taylor’s power law, it indicated that the sample size required for expanded leaves was the most. When there are 10 individuals on an expanded leaf, we need to sample 257, 649, and 209 leaves for Thripidae, Aphididae, and Aleyrodidae, respectively. When to estimate sucking-insects on expanded leaves, 123 expanded leaves have to be sampled based on 10 individuals per leaf. The results showed that the optimal sample size of the feeding guild was lower than single population. Because T. palmi was the key pest on eggplant, special attention was also paid to the effect of temperature on its life history traits and population parameters. Cohorts of T. palmi were reared on eggplant leaf at 15, 21, 25, 30 and 35℃ in growth chambers for three generations continuously. The results showed that survival rate of T. palmi from egg to adult was only 4 - 8% at 35℃, and the adult female and male lived only 2.9 and 2.8 days, respectively, without laying any eggs. Under the four temperatures the pre-adult stage took respectively 29.9, 19.6, 12.3, and 10.4 days to complete the development. The longevity of adult female became shorter as the rearing temperature increased, being 21.6, 20.2, 15.4 and 9.7days, respectively. The female had the highest fecundity (57.1 eggs/female) at 25℃, but its oviposition period was the shortest (ca. 21 days). A simple linear regression of developmental rate on temperature ranging from 15 to 30℃ showed that the lower developmental threshold (T0) was 7.7℃ (SE = 0.2℃) and the cumulated effective temperature ( K) was 227.2°D (SE = 3.3°D) for T. palmi to complete development from egg to adult. Based on the above value of T0 and K, and the 1999 to 2002 meteorological data of the Taiwan Agr. Res. Inst. at Wufong, we estimated that this thrips could complete 25 to 26 generations a year in central part of Taiwan. The results indicated that the age-specific fecundity ( mx ), the daily fecundity (fx5 ), and the age-specific maternity ( lxmx) was highest at 25℃. The intrinsic rate of increase (r) rose from 15, to 21 to 25℃, and fell at 30℃, being 0.033, 0.046, 0.157, and 0.118 day-1, respectively. The net reproductive rate (R0) was highest at 25℃ at 18.6 eggs. The mean generation time (T) shortened gradually from 15 to 30˚C; at 30℃ it was only one day shorter than 25℃. Consequently, we concluded that 25-30℃ is optimal for population growth of T. palmi on eggplant.