Study On The Field Investigation Technology Of Brontispa Longissima (Gestro)

Brontispa longissima (Gestro) is a new harmful insect invading in China,which feeds on the palmaceous plants. Through studing on the space distribution of B. longissima and clarified it's harm characteristics. accurate and effective sampling method was found out; Studies on the relationship between the damage of the coconut leavies and the population of B. longissima, simple and practicable classification and estimation models of population density were put forward. We found the foundation of scientific control B. longissima and releasing parasitic bee in proportion. The main results were as follows:1. The space distribution of B. longissimaIn horizontal distribution and vertical distribution ,the 5 aggregation indices is C>1,K>0,Ca>0,I>1,M*/M>1;lga>0 and b>1 in Taylor power equation and the F inspection value was remarkable by regression analysis with ovum,pupae and adult.β>1 in all the Iwao regression equation. butα>0 in the regression equation of larva,pupae and all the population withα<0 in regression equation of ovum,adult in vertical distribution;α>0 in the regression equation of ovum,larva,pupae,adult and all the population in horizontal distribution. In vertical distribution, the population assembling average(λ) >2 with larva and all the population andλ<2 with pupae,λ>2 in 2 fields andλ<2 in 2 fields with ovum and adult. In horizontal distribution,λ>2 with larva,adult and all the population,λ<2 with pupae, at the same timeλ<2 in 3 fields andλ>2 in 2 fields.The results showed an aggregation distribution pattern for the different stages of B. longissima in all the density with horizontal direction and vertical direction. And the level of aggregation of all the stages of B. longissima have the dependence on the density of population. The level of aggregation of ovum,pupae,adult went up along with the population density's increasing. In the vertical direction, the basic component of the larva,pupae and all the population distribution was the individual colony, but colonies attracted one another. They aggregate together someplace in the leaves. But the individuals interfered with each other while on the stages of ovum and adult, they distribute in the basal,middle and end of the leaves. In the horizontal direction, the basic component of all the stages of B. longissima distribution was the individual colony, but colonies attracted one another. In the vertical direction, pupae's aggregation was caused by some environmental factors,and the aggregation of larva,adult and all the population were caused by their behaviour and environmental conditions. For the ovum, the aggregation was caused by some environmental factors in Qionghai and Wenchang where the population density was low and the aggregation were caused by their behaviour and environmental conditions in Haikou and Sanya where the population density was high. In horizontal direction, the aggregation of larva and adult were caused by their behaviour and environmental conditions. The aggregation of pupae was caused by environmental factors. For the ovum, when the density of population is low in Qionghai,Wenchang,Sanya A and Sanya B, the aggregation were caused by some environmental factors; when the density of population was high in Haikou, the aggregation was caused by their behaviour and environmental conditions.2. Selection of representational injured leavesThe results of regression analysis between population amount (ovum,larva,pupae,adult) of every leaf with total population (ovum,larva,pupae,adult) of a coco in 4 fields of Qionghai,Wenchang,Haikou,Sanya showed that the F inspection value was extremely remarkable or remarkable by regression analysis between population amount of the second leaf with population amount of the whole coconut tree in all the fields. And the amount population of the second leaf is 72.00% of the total population of the whole coconut tree. So we could learn about the whole population through investigating the population amount of the second leaf. And much work load was avoided.3.Sampling method of B. longissimaWe investigated the B. longissima by 5 sampling method in every investigating field. All data of the investigation was dealt with by variance analysis way and t test, and we have got some results below in comparison with their absolute error,error rate and coefficient of variation. There are no difference between sample means from the fields by 5 sampling methods with population means. In 0.05 significance level, the results of t test between sample means of different stages by 5 sampling methods with population means showed that there was no significant difference .we made the whole population of every sample area as control, and calculated the means,absolute error,error rate and coefficient of variation between the results by every sampling methods with the control. then we averaged the absolute error,error rate and coefficient of variation of every field. We got the results as follows by comparing the precision of them. The absolute error of population means from low to high is interlacing sampling2/D²(61.978/M+0.4711),n= t²/D² (26.04096/M+3.96806) respectively which were studied by means of Iwao's regression model n=t²/D²[(α+1)/M+β-1] (t= probability assurance value;D= allowable error;M=average population). we made t equals to 1 and M equals to 10,50,100,150,200,500 respectively, when D equals to 0.1, The optimum sampling number of horizontal distribution is 667,171,109,88,78,60 respectively; The optimum sampling number of vertical distribution is 657,449,423,414,410,402 respectively. When D equals to 0.2, The optimum sampling number of horizontal distribution is 167,43,27,22,20,15 respectively; The optimum sampling number of vertical distribution is 164,112,106,104,102,101 respectively.5 The standard of the injury classification and forecasting model of population account for B. longissimaThe data of all the investigated fields were analysed, The exponential model of second leaf or second+third leaves assembly was chosen for trend forecasting model on the integration of results and practice. In all the models in regression analysis between the harming classification of 2 leaf or 2+3 leaves assembly with correspondence population, when we chosen the way of classification that 10% as start point and 20% as distance, the regression coefficients of power exponent model was high and they have only 5 levels. So it is of great value in practical. Regressive forecasting equation is y = 23.172e^0.7287x,y = 15.64e^0.8042x respectively(x= harm level).