ORGANIC MATTER DECOMPOSITION AND MICRO ARTHROPOD COMMUNITY STRUCTURE IN CORN FIELDS UNDER LOW INPUT AND INTENSIVE MANAGEMENT IN GUAÍRA ( SP )

A bstract: The ra te of organic m atter decomposition and the structure of the communities of m icroarthropods were com pared between two corn fields receiving contrasting agricultural m anagement practices (low input and intensive farm ing). The ra te of decomposition tended to be higher in the intensively managed field in the beginning of the growing season, bu t decreased to a level significantly lower than the observed in the low input field by the end of the growing season. This suggested that the biological community associated with the decomposition process could be negatively influenced in the intensively m anaged field. Analyses of the structure of m icroarthropod communities indicated differences between the two areas. The m icroarthropod populations present in the intensively m anaged field suffered ab rup t decrease in num bers as the season progressed. Key W ords: Soil fauna, organic m atter, decomposition, agroecosystems, farm ing systems, low-input agriculture


DECO M PO SIÇÃ O DA M A TÉRIA ORGÂNICA E ESTRUTURA DA COMUNIDADE DE M ICROARTRÓPODES EM M ILH O CULTIVADO DE FORM A INTENSIVA E COM BAIXO USO DE INSUMOS EM GUAÍRA (SP)
Resum o: A taxa de decom posição da m atéria orgânica e a e stru tu ra da com unidade de m icroartrópodes foram com paradas entre dois cam pos cultivados com m ilho m as recebendo m anejos distintos, sendo um cam po m anejado intensivam ente e outro com baixo uso de insum os.A taxa de decomposição foi m ais alta no cam po intensivam ente m anejado no início da c u ltu ra, m as decresceu p a ra um nível significativam ente inferior àquela observada no cam po com baixo uso de insum os ao final da estação.Tal tendência sugeriu que a e stru tu ra da com unidade dos organism os associados ao processo de decom posição poderia e star sendo negativam ente influenciada no cam po intensivam ente m anejado.Análises da e stru tu ra das com unidades de m icroartrópodes indicaram que diferentes com unidades estavam presentes nos dois cam pos.Âs populações de m icroartrópodes presentes no cam po sob m anejo intensivo sofreram queda a b ru p ta em núm eros, sendo praticam ente elim inadas já no segundo mês de desenvolvim ento da c u ltu ra.Descritores: F au n a do solo, m atéria orgânica, decom posição, agroecossistem as, m anejo agrícola

I n t r o d u c t i o n
The region of Guaíra, in the Northernmost part of São Paulo State, is characterized by a very intensive agriculture under rapid technological expansion, markedly with the employment of centerpivot irrigation system s.The heavy financial investment involved in such systems calls for very high yields from the irrigated areas, in order to repay the investment.Hence, allied to the high levels o f m ech an izatio n , the ag ric u ltu ral practices normally incorporate heavy inputs of fertilizers and pesticides.H ow ever econom ically sound this agricultural production system may appear, an evaluation of its environmental impacts is warranted (Abreu, 1994).
A preliminary qualitative evaluation o f the development o f some crops in several areas under intensive irrigated management indicated problems potentially linked to unbalances in the soil -plantsoil biota relationships.Firstly, large quantities of the organic residues from previous crops were still present on the soil in an alm ost intact state, suggesting that the decomposition process might be slow.Secondly, in spite of the intensive control effort, root pathogens were the main agricultural problem in most crops, possibly due to a relative lack of competition from antagonistic saprophytic competitors (Chung et a l , 1988).Lastly, it was possible to notice signs of toxicity in weeds and even crop plants, especially those sprouting at the bottom of furrows.Possibly, this was occurring as a result of fertilizers and pesticides being washed down and accumulating on these spots, due to excessive application.
As a result of these observations it has been hypothesized that unbalances in the dynamics of organic matter decomposition and in the structure of the soil community could be involved in those p ro b le m s.T his h y p o th e sis stem s from the observations of Santos & Whitford (1981) that soil microarthropods are responsible for a significant control of microbial populations, and consequently on the process o f decomposition.The role of microarthropods in microbial growth control is not a simple function of their numbers or abundance in the soil (which is greatly influenced by pesticide application regime), but results from a myriad of relationships among different trophic groups, spanning from saprophagous to fungivorous to nematode and mite predators (Argyropoulou et al. , 1993;Crossley et al., 1989;Moore, 1988).As a consequence of this intricate food-web and as a result of the very action of the microarthropods on the organic m atter and directly on m icrobial b io m a ss, the m ic ro b ial p o p u la tio n s can be maintained in the logarithmic phase of growth, improving the process of decomposition of organic matter, release of nutrients, and biological activity in th e so il(A n d ren eia /., 1988;Santos ei a /., 1981;Schroth et al. , 1992).
In order to address the basic hypothesis that unbalances in the community structure of soil m ic ro a rth ro p o d s and in the o rg an ic m atter decomposition process were being caused by the in te n siv e m anagem ent in irrig a te d are a s, a co m parison betw een an intensively m anaged irrigated field and a low input field was carried out (El Titi & Ipach, 1989).This paper describes the results o f such a comparison.

M aterial and M ethods
Study site: Two neighboring farms which applied different crop management practices were selected in a single soil patch in Guaíra.Two areas (A and B, low input and intensive agricultural management, respectively) were selected, each representing an experimental treatment.By the beginning of the siunmer growing season, both areas were sown with com.Area A was characterized by typical low input management, for the main economic activity of its owner was silkworm raising, and pesticides were not used in the property.Area B consisted of a plot under center-pivot, and the management involved frequent irrigation and heavy inputs of fertilizers and pesticides.
Assessment of organic m atter decomposition: In each of these areas a set of 18 litter bags (20 x 20 cm, 1 mm nylon screen, containing 5 g of dry mixed-leaf litter collected in a fallow field) were laid down under approximately 5 cm of soil between the crop rows.Two additional six litter bag sets were laid in each area one and two months after the first set, in order to permit monthly estimates in addition to the cumulative estimates provided by the samples initially laid.This sampling schedule is presented in Diagram 1.
After approximately 30, 60 and 90 days, a subset of 12 litter bags were randomly collected from each area (six for the monthly and six for the cumulative estimates) and brought to the laboratory.The content of each litter bag was put into modified Berlese furmels for mesofauna extraction, being the microarthropods fixed in ethanol-glycerin 3:1 (by volum e) m ixture.A fter com plete ex tractio n (a p p ro x im a tely 72 h o u rs) the c o lle c te d m icroarthropods were mounted in m icroscope slides, identified and counted.
After extraction, the organic material was dried at 105°C fo r 24 h o u rs, w eighted and incinerated at 600"C for 8 hours for the estimation o f org an ic m atter d ecom position (Santos & Whitford, 1981).In this method the decomposition rate is expressed directly as percent loss of organic carbon.These percentages were normalized by arcsine tran sfo rm atio n and the m eans w ere compared by Student's t tests.
Assessment of m icroarthropods: Microarthropod community was computed as the total number of each m icro arth ro p o d o rd er, total num ber of collembola and other arthropods, and total number of mites in each family.This arbitrary taxonomic subdivision was selected based on the role and im portance o f each group in organic m atter d e c o m p o sitio n and m ic ro flo ra -m e so fa u n a relationship, and position in the soil food web (Crossley et al. , 1989).
The total number of organisms per period per treatment, instead of mean number per sample unit, was computed due to the very well described patchy distribution of mesofauna in disturbed soils (Abbott & Crossley, 1982;Santos et al., 1978).The resulting data consisted then of counts of total m ic ro arth ro p o d num bers in each taxonom ic category, for independent samples in relation to fann (or agricultural management), for three distinct periods of time and for two periodical analysis, monthly and cumulative.As the interest of this smdy lays particularly on the effects o f agricultural m a n ag e m en t (low in p u t v e rsu s in te n siv e management), and not in the effects of time periods, the periods and periodical arrangement were studied independently, isolating only management effects in each comparison.The question to be posed in this comparison may be stated as follows: "Are relative numbers of microarthropod in each studied taxa independent of management practices?" It shall be noted that the question relates to relative numbers (evenness), meaning that even with accentuated drops in the total numbers (as it is likely to occur especially for the intensive management treatm en t, due to pesticide use) the relative proportions are sustained.This problem is typically one of comparison between observed and expected numbers, so the test extends naturally to this situ atio n (Snedecor & C o ch ran , 1967).The advantage of such a method is that it allows the construction of standardized deviations TABLES, so that each deviation (corresponding to each taxonomic category) may be assessed and discussed specifically.Statistically significant y} values in these TABLES indicate the instances in which the relative proportions of microarthropods differed in a given period, between the two treatments.
In order to address which individual taxa might have suffered deviations from expected numbers, sp e cific c o m p a riso n s w ere p e rfo rm e d in standardized deviations TABLES.The relative proportions o f m icroarthropods found in each treatment in each study period were applied in 12 X 2 (taxonomic categories x treatment) R x C TABLES (note that period 1 is the same for both periodic samples).Tests of significance for each comparison were performed for the following null hypothesis: Ho; The relative proportion of each microarthropod taxa is the same under low input and intensive agriculmral management.
The TABLES were constructed under the assu m p tio n th at the n u ll h y p o th e sis h o ld s, considering as expected values the total number of microarthropods in each taxa for both treatments, using the following expression: Standard deviations = Oij -Eij / SQRT Eij w here: O ij is th e o b se rv e d n u m b e r o f microarthropods in each taxa and Eij the expected number assuming Ho.For such a TABLE, a value o f ± 2 is co n sid ered to be asso ciated w ith meaningful deviations (Snedecor & Cochran, 1967)

Results and Discussion
The decomposition rates observed in this study for the summer growing season in cultivated com were very high in both treatments, reaching almost 70% loss of organic carbon in one single month.Such high decomposition rates are not uncommon for straw material in tropical situations (Lekha et al. , 1989).As can be observed in Figure 1, the rate of decomposition was significantly higher for the intensive agricultural management (area B) in the first month.This effect was probably related to the higher water (due to irrigation) and nutrient availability in this treatment for that period (Andren  et al. , 1995).In the second monthly period the two means were no longer different, while in the third month the samples from the low input treatment (a re a A) show ed a sig n ific a n tly h ig h e r decomposition rate.This result could be related to the increasing detrimental effects of the pesticides applied in the intensively managed field (Kajak, 1989).The same pattern was observed in the samples of the cumulative periods, with the rate of decom position being higher in the intensively managed field in the beginning, reversing to a higher rate for the low input field by the end of the study period (Figure 2).The number of microarthropods obtained in each treatm ent for each period o f tim e is presented in Figures 3 through 6.It should be noted that the total numbers of microarthropods are very d iffe re n t, w ith rem arkable d ecreases in the intensively managed field, again probably due to the detrimental effects of pesticides (El Titi & Ipach, 1989).T he re su ltin g v alu es o f fo r the comparisons of community structures between low input and intensive agricultural management for the five experimental periods (three monthly and two cumulative) is presented in TABLE 1.
The null hypothesis is rejected for the first two monthly periods as well as for the two months cumulative samples.By the end of the crop cycle the differences in relative proportions disappear, as numbers of microarthropods diminish abruptly especially in the intensively managed field.So, we conclude th at d iffe re n t soil m ic ro arth ro p o d community structures were associated with low input and intensive agricultural managements during the first two-thirds of the growing season.
A lth o u g h th is in fo rm atio n m ay be significant in agroecological terms, suggesting that impacts caused to the mesofauna by the agriculmral management can influence higher processes, such as organic matter decomposition, little is available for the understanding of implicit relationships that could be explanatory.In other words, what kinds of ecological unbalances could have been caused to the structure of microarthropod corrmiunities as to result in the observed results?Which taxa were influenced the most, and how so?
These questions were addressed by the analysis of specific standardized deviations in the observed numbers of microarthropods in relation to the numbers expected if the null hypothesis held.The analyses of these TABLES show that the most meaningful deviations occurred in the first period, at the time of crop emergence.The main contrasts were represented by a higher relative abundance of mites in the intensively managed field, while collembola were more important in the low input farming.Note that oribatid mites represented the prey in the food web of the high input agriculture (most probably with an important contribution of nematodes, not smdied), and that predacious mites (la e la p id a e , d ig a m aselid a e, cu n ax id ae, and imm ature gamasida) were proportionally more important.In the second monthly period, however, this pattern changed, with total numbers dropping bluntly in the intensively managed field.Predacious mites become rare in this field, and collembola still more important in the low input management.In the third monthly period most contrasts are no longer important, as the majority of groups were eliminated from the intensively managed field.
The same general pattern occurred in the cum ulative sam ples, even though in this case oribatid mites grew in relative importance in the low input field as the season progressed.Again, in the third period the differences disappear.This result may in fact be a bias resulting fi-om the skewed character in the data, caused by the extirpation of organisms in the later periods o f the intensive agricultural management treatment, causing the expected values to be closer to zero.These results show that while all groups were affected in the intensively managed field, the organism s closer to the base of the food web (especially collembola and oribatid mites) suffered the m ost.As it would be expected, w ith the elimination of the base of the trophic pyramid the whole community failed.
Even though the results obtained in this work permit only the drawing of conclusions in terms of association and not causation, since the data are essentially observational, the objectives p ro p o se d w ere ac h ie v ed .T he m ethod was sufficiently malleable to allow the assessment of specific relationships, with clear visualization of direction and degree of deviation, besides testing the hypothesis proposed.Finally, it was possible to obtain some insights for the design of further experiments to be conducted in order to show relationships of causation in soil microarthropod community structure and agricultural management.Two important conclusions can be obtained from this analysis: 1 -Manipulative and replicated experiments are needed to perm it inferences about causation relationships for agricultural management and microarthropod community changes in general, not only for the smdied situation.

Figure 1 -
Figure 1 -Organic matter decomposition rate in corn grown under low input and intensive management.Data points in each period showing different letters are significantly different at a = 0 .0 5 .

Figure 2 -Figure 3 -Figure 4 -NumbersFigure 5 -Figure 6 -
Figure 2 -Cumulative organic matter decomposition rate in com grown under low input and intensive management.Data points in each period showing different letters are significantly different at a = 0 .0 5 .

TABLE 1 -
Results of the five tests of significance performed for microarthropod community structure comparison between low input and intensive farming.

TABLE 2 -
Standardized deviations for microarthropod community structure relative to expected independence from agricultural management for three monthly periods.

TABLE 3 -
Standardized deviations for microarthropod community structure relative to expected independence from agricultural management for cumulative periods.