Systematic relationships of rice black bugs , S cotinophara spp . inferred using nonmetric multidimensional scaling technique and parsimony analysis

The systematic relationships of Philippine rice black bugs (RBBs), Scotinophara spp. was inferred using nonmetric multidimensional scaling technique (MDSCALE) and parsimony analysis to determine patterns of variation among the species and species groups. The MDSCALE result revealed a very distinct structure with clusters representing species groups visually apparent in the map. The distinct gaps in the phenetic spaces between species groups reflect the ease by which the species can be classified using numerical phenetics. On the contrary, parsimony analysis showed that the species groups did not form monophyletic groups. The results of the study are discussed in the light of patterns of differentiation among the species. It is suggested that more parsimony informative characters should be included to shed light into our basic understanding of the various evolutionary processes involved in the differentiation of these species of black bugs.

As early as 1982 to the present date, the rice black bugs (RBB) had been reported to destroy rice farming communities and many of the local government units in the entire Philippine Archipelago.Widespread yet difficult to observe, this highly invasive species has been found in the Philippines in four major islands to date: Palawan (1982), Mindanao (1992Mindanao ( -1997)), the Visayas (1998)(1999) (Miyamoto et. al., 1983, Barrion et al., 2008).Attraction of large numbers of bugs to lights is often mistakenly called outbreaks and result in anticipated infestation and yield loss. Rice black bugs rarely cause damage to the plant except in nonirrigated areas where no supplemental water is available for the rice plant.Hence, drought stress killed the plants hastened by the additional sap-sucking behavior of the rice black bugs (Heinrich, 1994).
Although only S. coarctata is commonly reported on rice throughout the Philippines, we suspected that a complex of related species exists because bugs from different localities behaved differently in feeding trials (Heinrichs et al., 1987).In addition, many of the bugs were found only on older plants and abandoned stubbles.
The most recent and comprehensive study of RBB in the Philippines placed Scotinophara species into four taxonomic groups: (a) tarsalis with a single species, (b) serrata with two species, (c) lurida with five species, and (d) the coarctata with 16 species (Barrion et al. 2007).Of the total taxa reported, 19 are newly described species.The new species belong to the serrata group.Two previously reported taxa, S. lurida (Burmeister) from Mindanao Island and S. ochracea (Distant) from Luzon Island were not examined due to lack of specimens from the current and loaned collections.
This study aimed to investigate the systematic relationships of the Philippine RBBs species belonging to the four taxonomic groups to determine patterns of morphological variation using the tools of phenetics and cladistics.The method of non-metric multidimensional scaling will be applied on the MDSCALE distance matrix to produce an unrooted tree, which will depict the phenetic relationships of the species.

MATERIALS AND METHODS
A total of 19 newly described species of black bugs from the Philippines were examined along with five previously known species: S. tarsalis, S. serrata, S. latiuscula, S. cinerea and S. coarctata (Table 1, Fig. 1) (Barrion et al., 2007).Central to problems with the determination of the phylogeny of organisms is the appropriate choice and application of outgroup data.For the purpose of this study, Antilochos nigripes of family Pyrrhocoridae which is a sister taxon to the Pentatomidae was chosen.
The 60 characters (16 binary and 44 multistate) identified from both sexes (Figs 2-4) were all unordered and equally weighted to refrain from any hypothesis about their transformation (Appendix A).In particular, coding does not imply whether a state is ancestral or derived.Unknown characters were coded as "?".The data matrix was subjected to non-metric multidimensional scaling analysis (MDSCALE, Kruskal, 1964).A minimum spanning tree (MST) was fitted to the data to determine the phenetic relationships of the species (Rohlf, 1970).This method was used to indicate the near neighbor distance relation among the points in the dimensional space.The graph was drawn using the program PAST v. 1.70 (Hammer et. al., 2001).The outgroup was excluded from the analysis because of its extreme separation from the in-group species based on the plot of the first two coordinates.
The data matrix was constructed and edited with WinClada as a shell program (Nixon, 1999) (Table 2).Parsimony analyses were conducted with Nona using the Parsimony ratchet (Goloboff, 1993).Two hundred iterations were performed (one tree per iteration), with the search repeated ten times.Tree characters were sampled for reweighing during the parsimony ratchet.All searches were done under the collapsing option "ambiguous" which collapsed every node with a minimum length of 0. Per character consistency (ci) and retention Systematic relationships of rice black bugs inferred using nonmetric multidimensional 115 indices (ri) were calculated to determine reliable and informative characters.Characters with consistency indices of 100 show no homoplasy.
Table1: Taxonomic groupings of the 24 species of previously (5) and newly (19)        Overall, the MDSCALE result revealed a very distinct structure with clusters representing species groups visually apparent in the map.Also, species in the same group are found near one another and linked through the MST.The distinct gaps in the phenetic spaces between species groups reflect the ease by which Scotinophara spp.can be classified using numerical methods.The species belonging to the Coarctata group are in the same region and are near neighbors of one another.These species have characteristically long anterior lateral margins in the pronotum the length of which is found to be approximately twice the entire length of the eye.Members of the Lurida group formed a connected set with members of this set shown as near neighbors of the Serrata group.The plot shows S. tarsalis as an isolated species and support the contention that this taxon be placed on its own monotypic species group.This species differs from the others in having a small body size, minute serrations on the lateral margins of the pronotum, anterior lateral margins in the pronotum which is oriented obliquely-upward, large and triangular anterior lateral spine, prehumeral spine shorter in length than the anterior lateral spine, and a widely or shallowly concave tip of the membranous conjunctiva.Moreover, S. pseudoserrata and S. serrata were the only members of Serrata species group.A distinct characteristic of these species is the presence of a spine in the pygophore (cross-section).
Analysis of the 60 characters identified 10 equally most-parsimonious trees (L = 12; CI = 58; RI = 78).Characters 10, 14, 17, 22, 23, 24, 25, 49, and 58 proved uninformative in the present dataset (See Appendix A for details).The strict consensus of these trees is presented in Figure 6 with areas of conflict in topology shown as polytomies.The presence of two polytomies was found in the basal as well as in the mid-apical nodes which may be the result of recurrent individual characters generated through different mechanisms (i.e.parallel evolution, homoplasy, convergent evolution).The results also show the lack of congruence between the results of the parsimony analysis and that of the MDSCALE presumably because of factors related to the choice of the outgroup, number of parsimony informative characters and the choice of the tree search algorithm.If character evolution for this group of species is to be founded solely on the strict consensus tree, then the observed similarities between species may have resulted from independent adaptations to a similar environment.Such recurrences in traits have been documented in groups in which most species are morphologically similar (i.e.sibling species) and especially when some of the species have invaded and have Systematic relationships of rice black bugs inferred using nonmetric multidimensional 123 adapted to the same selective environment.Thus, the existence of recurrent traits in the morphological data used in this study (as evidenced by the low per character ci values in Table (3), makes intuitive sense because morphological characters are said to interact with the environment more directly and frequently that other forms of data such as molecular characters, which are often assumed to be selectively neutral or nearly so (Kimura, 1983;Hedges and Sibley, 1994;Gillespie, 1991).Although, recurrent characters are important in testing hypothesis of adaptation in these species of black bugs, the relative extent of such recurrences/homoplastic gains of derived states may obstruct the resolution of phylogenies (Sytsma and Gottlieb, 1986;Gottlieb, 1988;Givnish and Sytsma, 1997).Future studies based on morphology must be conducted on a matrix containing more characters with greater degree of character consistency or at least with per character consistency indices (ci) greater than 50 (parsimony informative characters; Farris, 1989).
Character evolution and evolutionary diversification of the black bugs are discussed based on the strict consensus tree.Some of the characters were found to be sexually dimorphic in only a few numbers of species.These includes characters 12 & 13 (lengths of antennal segment I vs. II) in S. serrata, S. putikanica and S. trifurcata and characters 56 & 57 (length of the proboscis segments IV vs. III) in S. cinerea, S. pirurotonga, S. agusanortica.This phenomenon might represent examples of character evolution via a sensory exploitation system and divergence in feeding ecology (Ryan et al.1990 as cited by Penz and De Vries, 2002;Vincent, 2006).
A number of lineages are apparent in the strict consensus tree.Noteworthy is the group composed of the species S. luzonica, S. kalinga and S. arkwata, which forms one of the most basal lineages but is not supported by any synapomorphic characters.Few features may represent shared-derived characters within the in-group taxa.One synapomorphic character (length of antennal segment I greater than that of II in males) unites the species S. putikanica with the already known species S. tarsalis, S. latiuscula (Breddin) and S. coarctata (F.).The results of the analysis also identified a close relationship between the previously known species S. serrata and the newly described S. pseudoserrata and S. molavica.Morphologically, these species share the following characters: cross sectional position of pygophore subglobose without spine and length of basal theca and anterior theca subequal.
Several autapomorphies were mapped in the preferred cladogram, noteworthy of which are the unique features in S. tarsalis which includes the following: a slightly sinuate to sinuate lateral margin of the pronotum (6:0); presence of minute serration on the lateral margins of the pronotum (2:0); anterior lateral margin pointed in an oblique-upward position (3:1); presence of a large triangular anterior lateral spine (10:2); length of the prehumeral spine less than that of the anterior lateral spine (11:1); presence of two spermathecal bulb processes (46:2) and presence of an exserted Paratergite IX (49:1).Almost all of these characters were also identified in the phenetic analysis.
S. serrata differs from the others species in having four autapomorphies, which includes: a slightly cleft tip of the scutellum, tapered separated tips of the penis, slightly exserted ejaculatory duct, and the presence of three longitudinal veins below the discal cell.Two other species are defined by a single autapomorphic character each.A truncate to slightly emarginate tip of the scutellum is autapomorphic for S. arkwata while an emarginate tip of the scutellum is unique for S. molavica.
Mapping of character state distributions across the preferred cladogram was done only on characters that exhibited a certain degree of homoplasy (ci values of 50 and above) (Fig. 7).The trees produced permit the study of character evolution in these species of bugs.
Among the characters that appeared twice in the preferred cladogram is the presence of a head notch in S. tarsalis and S. serrata (7:1), both of which belong to separate lineages.Only one character showed a high consistency index of 100 (character 12: length of antennal segment I vs. II in males), the character states of which grouped the species into three clusters: the first composed of S. kabangkalanensis, S. luzonica, S. kalinga and S. arkwata; the second composed of S. putikanica, S. tarsalis, S. latiuscula (Breddin) and S. coarctata (F.); and the last group comprised of all other species.Results showed that the three species are known to be pests of agricultural crops which include S. latiuscula (Breddin), S. coarctata (F.) and S. cinerea (Le Guillou) (59:1).The first two are shown to be more closely related in the cladogram produced.However, this study is taxonomic, and not designed to detect pest status.

CONCLUSION AND RECOMMENDATION
The genus Scotinophara spp. is probably among the most economically important taxa in the rice agroecosystem.Because of this, it has been the object of several morphological studies which have revealed it to be a complex of species in the Philippines.Thus, there are probably more than one species causing damage to rice in the islands of Mindanao, Siquijor, Negros, Panay, Bohol, and Luzon in the Philippines.The 14 new species of rice black bug belonging to the coarctata group recently described where 4 new species belong to the lurida group and a single taxon to the serrata group are considered new to science.Practical consequences of proper species identification at the species level may affect the design and implementation of quarantine restrictions, management and control programs.
Therefore, this study therefore investigated the systematic relationships among the RBB species in the Philippines using the tools of phenetics and cladistics and determined patterns of variation among the species groups.Results of the phenetic analysis using MDSCALE showed that the species groups were distinct and that the members of each group were similar to one another.Few morphological differences distinguished species among the groups.Overall, the plot of the minimum spanning tree showed a stable classification of the RBB into four species groups.The species within each group were connected, respectively.
The ten equally parsimonious trees resulted in a consensus tree of 12 steps and CI of 58.Of the 60 characters used in the analysis, nine proved uninformative.Inspection of the per-character ci should have a high level of homoplasy in the current data matrix.The results of the parsimony analysis did not support the contention that the species belong to four groups previously identified based on the MDSCALE plot.A number of monophyletic groups are apparent in the strict consensus tree.Six of the Mark Anthony J. Torres et al. 130 species scored are autapomorphic species supported by few unique character states.Noteworthy, the character states diagnostic for S. tarsalis, S. arkwata, S. molavica, S. mlanga, S. latiuscula (Breddin), S. ilonga, S. pirurotonga and S. alegria.
This study is far from being complete.While the true nature of the relationships of these rice black bugs are beginning to unfold, there are other traits that need to be explored to shed light to our basic understanding of the various evolutionary processes involved in the differentiation of these species of black bugs.Molecular data, behavior, feeding ecology, embryology and more informative morphological characters might generate new information that will be very useful in establishing the relationships of the species.

Fig. 6 :
Fig.6: Single cladogram found by Nona based on qualitative characters from species of black bugs from the Philippines.Closed circles represent synapomorphies, open circles represent reversals.Numbers above circles are character numbers (corresponding to Appendix 1); numbers below circles are states of the characters.

Table 3 :
Length, consistency and retention indices of the characters included in the study (NA -Uninformative characters).