The Vertical Distribution of Helopeltis bradyi and Oxyopes javanus on Tea

Helopeltis bradyi is the main pest of tea plants. Ecological characteristics of this pest are important to be understood to support the development of their management and control measures. This study aimed to determine the coexistence and vertical distribution pattern of H. bradyi and its predator, Oxyopes javanus, on tea plant parts. The study was conducted at the PT Pagilaran tea plantation in Central Java, in the 2018 rainy season. Population observations were carried out in situ on 20 infested sample-trees taken randomly, for 10 consecutive days, in the morning, at noon, and in the afternoon. Vertical distribution patterns were determined based on Poisson dispersion index (DI), negative binomial, and Green index (GI). The results showed that the O. javanus spider was found preying on H. bradyi. Coexistence between this pest and predator in the same part of the plant, the pest, and the predator, occurred in the morning were 50.0, 42.8, and 7.2%; at noon were 58.3, 41.7, and 0%; and at the afternoon were 66.7, 33.3, and 0%, respectively. The parts of the plant for the coexistence are pekoe leaves; the 1st, 2nd, 3rd, and 4th of young leaves; and 1st of older leaves. The pattern of vertical distribution in the morning, at noon and in the afternoon for H. bradyi was the weak clump, while for O. javanus was uniform. The ratio of predator: prey in the morning, at noon and in the afternoon was 1:10.7, 1:16.7, and 1:10.0, respectively.


INTRODUCTION
In 2013, Indonesia ranked 6th as the world's largest tea producing country with a production of 150,000 tons per year (Ramhot, 2015). However, tea production has been decreasing per year (Pusdatin, 2015). One of the causes is the presence of pests on tea plantations. Helopeltis was originally a minor pest in tea plantations, has been a major pest (Anonymous, 2018). The yield loss reaches 40% and the predicted loss is 50-100% (Gusti & Soesanthy, 2014). Integrated control techniques are needed to control Helopeltis, hence minimize the use of conventional pesticides applied to tea plants. Mechanical control and the use of natural enemies are the control techniques for pests of the tea plant, including Helopeltis (Hazarika et al., 2009). There are various kinds of natural enemies of Helopeltis on tea plantations, for example, predatory arthropods. The spiders as predators dominate (43%) compared to other predators. One of the genera found is Oxyopes (Das et al., 2015).
The pest distribution pattern has been widely investigated as basic information in the study of ecological characteristics and pest control management (Muraleedharan et al., 1988;Wagiman et al., 1998;Siswanto et al., 2008). The vertical distribution pattern of pests is important to be studied because it can affect sampling programs, describe the condition of the population, and can be used to analyze predator-prey relationships (Southwood, 1978). Predator-prey relationships are very important in agroecosystems. The characteristics of predatory behavior to the distribution of prey can be studied vertically in host plants (Wagiman et al., 1998). According to Siswanto et al. (2008) (Basnet & Mukhopadhyay, 2014). Therefore, this study aimed to determine the coexistence and vertical distribution pattern of H. bradyi and O. javanus in the tea plant.
This research provided basic information in an effort to control H. bradyi, both using mechanical technique and natural enemies.

Research Location
This research was conducted at the PT Pagilaran, Batang, Central Java tea plantation in April 2018, by monitoring the population of H. bradyi and O. javanus. The type of tea plant in PT Pagilaran was based on age criteria: producing plants, young plants not yet produced, and young plants produce. Young plantations produce located at an altitude of 720 masl and covering an area of 4 ha were selected in this study as observation sites. This selection was based on plant structure influenced by the age of the plant and the year of crop cutting. The criteria for selected young plants were plants with the age of 10 years in the plucking status of the Jendang (first cutting year). This would make the plant structure has not many branches, twigs, and leaves ( Figure 1). Based on these criteria, the condition of tea plants becomes easier to be observed visually. The sample selection was based on the existence of H. bradyi attack on selected tea plants. Preliminary observations were made to determine the optimal n value in sampling.

Sampling Technique
Observations were carried out visually and in situ by observed the existence of various stages of H. bradyi and O. javanus in parts of tea plants (Figure 1), which was calculated based on the number of individuals observed for each stage: eggs, nymphs, and adults were calculated as one individual. In the egg stage with the appearance of a pair of the chorion, strands were counted as one individual. If egg mass was found, then counted the number of chorion strands as the number of eggs. The number of the nymph and adult stage was calculated more easily than eggs.
Samples were taken purposively by selecting 20 plants attacked by H. bradyi and carried out randomly. The purposive mechanism for random sampling was that there are criteria for selecting samples (plants attacked by H. bradyi), hence the sampling was done purposively. If selected plants did not meet the criteria, the closest young plants that meet the criteria from selected plants were selected randomly. Observations were conducted on three categories of time: in the morning at 06.00-08.00 (Western Indonesian Time), in the noon at 11.00-13.00 (Western Indonesian Time), and in the afternoon at 16.00-18.00 (Western Indonesian Time). Ten replications were used for each observation time on a different day, without a rainy day. The sample unit used was one tea plant. Determination of optimal n was employed to determine the size of sampling. On the observing vertical distribution with purposive sampling, the character type of the sample is homogeneous thus to obtain the optimal n value using this following formula (Southwood, 1978): [1] n = number of samples, S = standard deviation, x = average population, E = standard error. Based on the calculation of preliminary observation, the optimal n value was 20 units with a standard error of 0.16%. The parameters observed were the number of H. bradyi in various stages (eggs, nymphs, and adults), and the number of O. javanus.

Data Analysis
Data of the vertical distribution of H. bradyi and O. javanus was analyzed based on the Poisson dispersion index (DI), negative binomial (k ), and Green Index (GI) by Ludwig and Reynold (1998), reported that insects will form a spatial distribution pattern clump naturally. Siswanto et al., (2008) also stated that Helopeltis has a clump distribution pattern in cashew plantations. Testing of this distribution pattern was through several stages: Ho was rejected from the Poisson distribution thus the insect population was not randomly distributed, the negative binomial test was used to justify if the targeted insect population was clump distributed, and the Green index was used as a further test to determine the grouping degree of the target insect population. The three analysis was needed to conclude comprehensively and mutually as described by Ludwig and Reynold (1998).
The formula for calculating the Poisson dispersion index (Ludwig & Reynold, 1998): The Poisson dispersion index (DI) was interpreted in the following categories: if the value of DI = 1 was random spatial distribution; DI < 1 was uniform; and if DI > 1 was clump distribution. The value of d (samples ≥ 30) or Chi-Square 2 (samples < 30) was used to test Poisson model.
The formula for calculating a negative binomial was k, if the average population is small (less than 4) then the formula used is (Ludwig & Reynold, 1998): k = log 10 (n/n0) = k log 10 [1 + (x ̅/k)] [5] k = negative binomial index, x ̅ = average of data sampling, n = total data, and n0 = number of data with population value of 0 (zero).
The formula for calculating the Green Index (GI) (Ludwig & Reynold, 1998): GI = Green index, x ̅ = average data sampling, s 2 = variant, n = total data. Based on Ludwig and Reynold (1998), a range of values 0-1was used to read the Green index value, where the value of 0 indicated that the population is random and values close to 1 or 1 indicated that the population is grouped with increasingly strong. Based on this finding, the degree of population grouping of H.
Bradyi was in Table 3. to read it easier.

Coexistence of Helopeltis bradyi and Oxyopes javanus
The H. bradyi was found in the internode of 1st, 3rd, 4th, 5th, 5th, 6th, 7th, twig, and stem ( Figure 4). The existence of pests without their natural enemy showed the displacement behavior of H. bradyi, which was moving towards the rootstock to hide when the disturbance was present (Roy et al., 2015). Meanwhile, the existence of O. javanus on the part of the tea plant was slightly found and only in the third part of the older leaves. This showed that O. javanus is a predator whose existence following the existence of their prey.  (Symondson et al., 2002).

Vertical distribution of Helopeltis bradyi and Oxyopes javanus
The distribution of H. bradyi and O. javanus in the tea plant was showed in Table 3 and  (Southwood, 1978;Ludwig & Reynold, 1998;Newman, 2013). The behavior of a grouping of insects could be interpreted that there are obstacles to the population of insect existence. Grouping behavior is that individuals group in parts preferred in their habitat, the existence of environmental plurality or reproductive models (Quinn & Dunham, 1983 (Bhuyan & Bhattacharya, 2006). Roy et al (2015) also explained that after hatching, nymphs immediately begin to eat young leaves and shoots.
The vertical distribution pattern of O. javanus was uniform because it has an DI < 1. The pattern of the uniform distribution is a pattern rarely found in organisms, hence the statistic analysis for that pattern is unavailable yet (Southwood, 1978;Ludwig & Reynold, 1998). Nature is a multifactorial system, where there are many interactions (biotic and abiotic) which can stimulate the existence of a pattern (Quinn & Dunham, 1983cit. Southwood, 1978. In this study, the distribution pattern of O. javanus was concluded based on the rejection of the statistical test of O. javanus population in the Poisson test (random distribution), negative binomial test, and Green Index (clumped distribution) (Table 3). Newman (2013) reported that the conclusion of the distribution pattern for population data rejected in the Poisson and binomial negative tests are categorized into a uniform distribution. Furthermore, uniform distribution patterns are the result of negative interactions between individuals, such as competition in feeding preference or problems of separation (Quinn & Dunham, 1983cit. Southwood, 1978, and this could occur to various types of organisms, including O. javanus. In general, the grouping pattern of H. bradyi and O. javanus in the morning, noon, and afternoon has the same pattern ( Figure 6). in the afternoon of 66.7%. The coexistence was in pekoe leaves, the 1st, 2nd, 3rd, 4th, 4th, and 1st of young leaves. Therefore, when the prediction of H. bradyi attack was high, scissor system and plucking it once out without leaving shoots in the field was recommended as control techniques.
Furthermore, besides reducing the pest population on the shoots, also reducing the food availability for the eggs hatched after the plucking. O. javanus population in the field could be increased by a mass rearing (augmentation) to maximize the potential for predation on H. bradyi.     (Ludwig & Reynold, 1998). c. Poisson test (random) with N ≥ 30 used value of d; If d > 1.96, Ho (random) was rejected, which did not follow the Poisson Distribution or not a random distribution d. Negative binomial (̂) is a grouping parameter with a range of 0 -2, the higher of the value of ̂ until reaches an infinite number showed an approach to random distribution (Southwood, 1978) e. Green Index (GI) showed the degree of grouping population (Tabel 3). f. *The uniform distribution pattern was obtained if Ho was rejected in the Poisson Distribution and Negative Binomial resulting in negative values (Newman, 2013) Ludwig & Reynold (1998)