The effect of starvation on the metabolic rate and microanatomy of Galumna elimata ( Acari : Oribatida )

The physiological parameters mortality, mass, oxygen consumption and amylase activity, and microanatomical features of the digestive tract, mesenchym and reproductive organs were used to characterise starvation in Galumna elimata. The mites were reared in sterilised plastic vials containing moistened zeolite at 25 °C and a 12:12 photoperiod. The control group was kept under the same conditions, but pieces of bark covered with the green bark alga, Desmococcus vulgaris (syn. Protococcus viridis), were added as food for the mites. The physiological parameters were recorded after 21 days, and the microanatomical after 21 and 42 days. The guts of the starved mites were empty or filled with mucoid substances, while the guts of control mites contained food boli formed from algal cells. The mortality was significantly higher in starved animals. The mortality after 42 days of starvation was higher in males than females. The fresh mass of starved individuals significantly decreased while the water proportion content of their body tissues increased. Oxygen consumption of the starved mites was lower. Starvation did not influence the activity of amylase. Glycogeneous granulae were characteristically absent, and mucoid substances present in the guts of mites starved for 21 days. The activity of mesenteral and caecal cells, proventricular glands and cells of salivary glands was reduced after 42 days of starvation. The cells of the seminal vesicles were reduced and contained no spermatic cells in males starved for 42 days. Starved females probably resorbed immature oocytes, but had eggs in their oviducts. Starvation induces ovovivipary or larvipary in Galumna elimata. Dijkman, 1993; Urbášek & Starý, 1994; Šustr & Starý, 1998). It is not always easy to interpret these results because: (i) The difficulty of establishing outdoor conditions indoors (Woodring & Cook, 1962a). (ii) Food preference tests are influenced by the spectrum of diets offered. Galumna elimata preferred the green bark alga, Desmococcus vulgaris, in the laboratory but this species is not present in soil (Hubert et al., 1999; Hubert & Lukešová, 2001). (iii) The quantification o f food consumption (Saichuae et al., 1972) or defecation (Hubert et al., 1998) does not provide information about assimilation efficiency. Scheloribates laevigatus has a high defecation rate when fed Desmococcus algae and Penicillium spores, but the spores and some o f the algal cells pass through the gut undamaged (Hubert et al., 1999; Hubert & Lukešová, 2001). A higher food intake could compen­ sate for energetically poor diets. (iv) The presence o f substrate specific enzyme activity indicates, but is not proof o f substrate utilisation. For example, Achipteria coleoptrata and Scheloribates lae­ vigatus differ in chitinase activity (Siepel & RuiterDijkman, 1993), but neither species digest fungal cell walls, where chitin is the main compound (Hubert, in prep.). The physical and chemical conditions in the oribatid gut are unknown. Enzyme activity, measured in vitro, does not correspond to the amount o f the sub­ strate digested in vivo. There are indications that there are qualitative differences in the chitinase activity in two populations o f Steganacarus magnus (see Smrž, INTRODUCTION The feeding biology o f oribatid mites has been inten­ sively studied, especially the interactions between oribatids and the micro-organisms responsible for plant litter decomposition (reviewed by Wallwork 1983 and Seastedt 1984). In order to achieve a better understanding of the role of oribatids in soil systems and litter decompo­ sition, several authors have described their feeding habits and assigned them to feeding guilds (Schuster, 1956; Luxton, 1972; Kaneko, 1988; Siepel & Ruiter-Dijkman, 1993). However, the food resources used by oribatids in soil are difficult to investigate due to the heterogeneity of soil systems. Many experiments and observations have been made on their feeding habits under laboratory conditions. In addi­ tion such information was also obtained from rearing experiments (Cleat, 1952; Sengbush, 1954; Woodring & Cook, 1962a; Stefaniak & Seniczak, 1976). Food prefer­ ence tests were used by Wallwork (1958), Hartenstein (1962), Luxton (1972), Trávníček (1989), Siepel (1990), Rihani et al. (1995), and Maraun et al. (1998). The micro­ organisms have been identified and the structural changes that occurs in the food during its passage through the gut have been determined by dissection and the cultivation of gut contents (Schuster, 1956; Hoebel-Mavers, 1967; Behan & Hill, 1978; Behan-Pelletier & Hill, 1983; Smrž, 1992a; Smrž, 1996). The presence o f digestive enzymes gives some information on the compounds they utilise (Zinkler, 1971 and 1972; Luxton, 1972; Dinsdale, 1974; Zinkler et al., 1986; Siepel, 1990; Siepel & Ruiter-


INTRODUCTION
The feeding biology o f oribatid mites has been inten sively studied, especially the interactions between oribatids and the micro-organisms responsible for plant litter decomposition (reviewed by Wallwork 1983 andSeastedt 1984).In order to achieve a better understanding o f the role o f oribatids in soil systems and litter decompo sition, several authors have described their feeding habits and assigned them to feeding guilds (Schuster, 1956;Luxton, 1972;Kaneko, 1988;Siepel & Ruiter-Dijkman, 1993).However, the food resources used by oribatids in soil are difficult to investigate due to the heterogeneity o f soil systems.
Many experiments and observations have been made on their feeding habits under laboratory conditions.In addi tion such information was also obtained from rearing experiments (Cleat, 1952;Sengbush, 1954;Woodring & Cook, 1962a;Stefaniak & Seniczak, 1976).Food prefer ence tests were used by Wallwork (1958), Hartenstein (1962), Luxton (1972), Trávníček (1989), Siepel (1990), Rihani et al. (1995), and Maraun et al. (1998).The micro organisms have been identified and the structural changes that occurs in the food during its passage through the gut have been determined by dissection and the cultivation o f gut contents (Schuster, 1956;Hoebel-Mavers, 1967;Behan & Hill, 1978;Behan-Pelletier & Hill, 1983;Smrž, 1992a;Smrž, 1996).The presence o f digestive enzymes gives some information on the compounds they utilise (Zinkler, 1971 and1972;Luxton, 1972;Dinsdale, 1974;Zinkler et al., 1986;Siepel, 1990;Siepel & Ruiter-1998).Scheloribates laevigatus mites reared in the laboratory differ from those collected in the field in their amylase activity (Hubert et al., 1999).The amy lase activity o f Galumna elimata changed due to thermal acclimation (Sustr & Hubert, 1999).Such intra specific changes in enzyme activity may be due to either a nutritionally poor diet or changes in tempera ture and therefore subject to misinterpretation.Generally, the possibility o f compensatory o f digestive activity (at the level o f food consumption as well as con centration o f digestive enzymes) makes the differences between species difficult to interpret.Neither the most attractive (preferred in the laboratory tests) food nor the food consumed or defecated at the highest rate can be assessed as the most nutritionally suitable without knowing the efficiency with which it is assimilated.Even specific substrate enzyme activity is ambiguous if there is knowledge o f intra-species variability and compensatory ability.
These problems are resolvable if very long term experi ments are used in which the survival and reproductive success o f a species feed a particular food are monitored.Another approach is to use more sensitive parameters o f the energy budget in standard feeding experiments, and experimental investigations o f the variability and physio logical adaptations o f oribatid digestion.Starvation repre sents an extreme feeding regime and could provide a useful model for such investigations.
The aim o f this study is to describe the oribatid Galumna elimata tolerance o f starvation, and the changes in body water content, respiration and digestive enzymes, and microanatomy o f the digestive tract, mesenchyme, and reproductive organs that occur in this species during starvation.The objectives were to find physiological and microanatomical features that indicate starvation or a low energy intake.

MATERIAL AND METHODS
Galumna elimata (C.L. Koch, 1841) is a common oribatid mite in meadows in Central Bohemia (Czech Republic).The specimens originated from meadow soil collected near the centre of Říčany town, 20 km east from Prague, 406 m. a. s. l..The samples were collected in October and November 1999.Mites were extracted using modified Berlese-Tullgren apparatus (extraction temperature was 35°C, water was as the collecting fluid and was replaced daily for 5 days).The adults were kept in plastic vials (volume 250 ml), the bottoms of which contained plaster of Paris.The vials were stored in a refrigerator (at 8°C) for one week before the start of the experiments.
Two groups of oribatids were established: starved mites and a control group fed on the green bark algae, Desmococcus vul garis (syn Protococcus viridis).Galumna mites are able to com plete their life cycle on this food (Sengbusch, 1954).Algae is a more attractive food for Galumna elimata than plant litter (Hubert, in prep.).Both groups were reared in plastic vials con taining 15 g of sterilised zeolite (Chemko®, Slovakia) moistened with 10 ml of sterilised distilled water, and remoistened weekly.The vials were kept in a controlled regime (12L : 12D hours, temperature 25 °C). Experiments: (1) Mortality.Both the starved and the control groups consisted of 6 rearing chambers, each represented by a vial with 18 individuals.After 21 and 42 days these individuals were placed into Petri dishes, the bottoms of which were lined with moistened filter paper, and observed.A circle (diameter 1cm) was drawn on the filter paper and the mites placed in this circle.The mites found in the circle after twelve hours were classified as dead.The differences in mortality of the control and starved groups were tested by log linear analysis (weighted by the number of individuals in each group) within Splus®.
(2) Mass of individuals.The living individuals were weighed (one sample usually con tained 16 individuals) on a microbalance (R160P -Sartorius®; accurate to 0.01 mg).The six replicates were per experimental group.Samples were dried at 110°C for 4 hours and weighed again to obtain the dry mass.Water content was calculated as the difference between fresh and dry mass.The mass differences were tested using the non-parametric Kruskal-Wallis test (Stat-graphics®).
(3) Respiration.A modification of the manovolumetric respirometer was used for the respirometric measurements.Verdier (1983) described its theory, based on the differential equation of the gas law.The length of the fluid column in a glass capillary of a given cross sectional area depends on the pressure and volume changes occurring thin the respirometric cell.One millimetre of dis placement corresponds to 0.01 pl.The apparatus consists of a thermostatically controlled airtight water bath with several res pirometric units almost completely submerged in the water.Every respirometric unit consists of glass capillary (internal diameter about 0.2 mm, 30 cm in length).The upper end of the capillary leads into the reservoir of manovolumetric fluid (xylene), which is open to the internal air of the bath.The bottom (submerged) end of the capillary leads into the respiro metric cell.The respirometric cell is divided by a fine net into a space for the animal and one for the CO2 absorbent (10 pl of 0.1N KOH).One circle of filter paper (3 mm diameter), wetted with distilled water, was used to maintain a high humidity in the respirometric cell.
Each respirometric unit was loaded with KOH, a moist piece of filter paper and test animals (16 individuals) seated and then these units plus several control ones (without animals), were incubated for 30 minutes in the bath before the measurements began.Then the reservoirs of the units were quickly filled with 2 ml of the manovolumetric fluid, and the bath was hermetically closed.The start position of the meniscus of manovolumetric fluid was recorded after 5 minutes, and then regularly at 20 minute intervals.The respiration rate was measured at a con stant temperature (20 ± 0.02°C) for 2 hours.The mites were weighed on a R 160P microbalance (Sartorius) immediately after the measurement.Oxygen consumption rate (M) was cal culated as M = s.dl.(K +K'.L), where K and K'.L are experi mental data, which take into account initial parameters such as pressure, volume and temperature, s is the inner diameter of the capillary tube and dl is the change in the length of a fluid column in a tube.Respiration rate was expressed in live-mass specific units (ml O2. g-1.h-1) of the starved (7 replicates) and control (9 replicates) groups after 21 days.Non-parametric Kruskal-Wallis test (Statgraphics®) was used to evaluate the differences between the groups.
(4) Activity of digestive enzymes.Amylolytic (EC 3.2.1.1)activity was assayed after 21 day.Whole-body homogenates of 64 living mites were used, with 12 replicates per group.The mites were weighed and homogenised in 2.5 ml of phosphate (Britton-Robinson) buffer (pH 7) and centrifuged (6000 rpm, 7 min).The pH optimum of Galumna elimata amylase was 7 (Sustr, unpublished).The S-Test using specific chromolytic substrates (Institute of Chemistry Brati slava, Slovakia) was used for the amylase assays (see Sustr & Hubert 1999).The catalytic activities of the enzymes were expressed in mg of decomposed substrate per hour per 1 g of fresh oribatid body mass (mg.h-1.g-1).The significance of the difference between experimental groups was tested using non parametric Kruskal-Wallis test (Statgraphics®).
(5) Microanatomical observations.Living individuals from both groups were fixed in modified Bouin-Dubosque-Brasil fluid (Smr □, 1989) after 21 and 42 days.The fixed mites were embedded in paraplast, sectioned (thickness 5-7 pm), and stained in Masson's triple stain.8) or empty (Fig. 6).After 42 days of starvation their tissues appeared reduced (Tab.2, Fig. 6).Some fluid substances present in the sali vary glands (Fig. 19), and mucoid substances in the caeca (Fig. 5) and by the mesenteral cells (Fig. 10).These sub stances were mixed with the ingested algal cells in fed individuals, or concentrated in the middle of mesenteron in starved individuals.In starved animals, the mucoid substances filled the whole mesenteron (Fig. 8) and more concentrated mucoid droplets formed boli (Fig. 9).The boli were passed through the gut into the rectum (Fig. 18).There were no structural differences in the mesenteral and faecal boli.mites.The cells were filled with dark stained granules (Fig. 10).There were green stained microvilli on the apical parts of the cells.The nuclei were relative large and well stained.These cells produced little apocrine secretion compared to the control specimens, there were fever granulae in the cells, and the thickness of mesenteral cells was not reduced after 21 days of starvation (Fig. 11).After 42 days, the mesenteral cells were very thin, and the nuclei still lacked dark stained granulae.No microvilli were observed (Fig. 12).The caecal cells in control individuals exhibited intense apocrine secretion (Fig. 13).The cells were strongly vacuolized and pro duced mucoid substances and large green stained spherical granulae.The caeca o f mites starved for 21 days were similar to those of the control group, only lacking green granulae (Fig. 14).After 42 days o f starvation, the cells o f caeca were thinner and poorly vacuolized and lacked apocrine secretion (Fig. 18).

The
The cells o f the colon (Fig. 8) were similar to those of the model species (see Woodring & Cook, 1962b;Hoebel-Mavers, 1967).
The rectum o f control individuals had a reduced brush border (Fig. 16), while the rectum o f animals starved for 21 days had a well-developed brush border (Fig. 17 The mesenchymal cells o f control specimens contained glycogeneous granulae (Fig. 15).These glycogeneous granulae were not observed in starved mites, except for one gravid female that had been starved for 21 days.Extra-intestinal bacteria were observed in one specimen starved for 21 days (Fig. 21).
The male/female ratio tended to 1:1, however females prevailed over males after 42 days o f starvation.Testes and seminal vesicles were not different from those o f the model species Ceratozetes cisalpinus (see  -Galumna elimata, control group -sagital section; 5 -horizontal section; 6 -horizontal section of an individual starved for 42 days; 7 -intensive consumption of algae by control individual; 8 -mesenteron of an individual starved for 21 days filled with mucoid substances of; 9 -mucoid bolus in an individual starved for 21 days; 10 -mesenteron of a control individual, the arrows point to mucoid package of food; 11-apocrine secretion in the mesenteron of an individual starved for 21 days; 12 -mesen teron of an individual starved for 42 days.Abbreviations used: c -colon, ca -caecum, cm -cheliceral muscles, cu -cuticle, eegg, fb -food bolus, m -mesenteron, mc -mesenteral cell, md -mucoid droplets, ms -mucoid substances, mu -muscles, ooesophagus, ph -pharynx, r -rectum, sg -salivary glands, syn -synganglion.Scales: 0.1 mm ... 4-6; 0.5 mm ."7-9;0.025 mm...10-12.1992a).As in the model species Trichoribates trimacu latus (see Smrž, 1992a)  The difference in the rectal brush border between starving and fed individuals is related to water absorption (see Smrž, 1992a).The passing o f an algal food bolus requires little time, and this food probably provided Galumna elimata with water, so the brush border could be reduced.A similar reduction is observed in Schelo ribates laevigatus fed on algal diets (Hubert & Lukešová, 2001) .The well developed brush border present after 21 days o f starvation may be related to effective water absorption.However, the higher mean water content o f starved mites contradicts to this explanation.All absorp tion mechanisms (including water absorption) might show on increase in efficiency in the initial period o f star vation and result in the higher water content in starved mites.The efficiency o f absorption may be reduced again later in order to save energy.

69Z
All other changes may be late indicators o f starvation, which appear when approaching the LTs0 values.After 42 days o f starvation, the reduction in cell contents and cell granulation in the mesenteron, colon and rectum were marked.These microanatomical features can be recom mended for identifying starvation in long term experi ments.
Although the production o f mucoid substances in starved mites decreased, some starved individuals pro duced mucoid boli.This indicates the starved individuals produced the same substances as fed individuals (also see Sustr & Hubert, 1999) They are probably partly reabsorbed oocytes, or aggre gated hemocytes (cf.Smrž, 199S).This phenomenon is not reported for natural populations o f mites (see Smrž, 1992a).The absence ofjuveniles as than starved indicates that there probably was no oviposition.In the laboratory.Scheloribates laevigatus females exhibit ovovivipary when fed filter paper, but the larvae died after a few days (Hubert, unpublished).The higher number o f germ cells in the embryos o f starved Galumna elimata females indi cated ovovivipary.Larvipary and ovovivipary is reported in some oribatid mites (Wallwork, 1967;Webb & Elmes, 1979;Hubert, 2000).Starved females o f Galumna did not deposit eggs.In the soil, although females o f most species have eggs in their oviducts throughout the year, oviposi tion occurred only after an environmental change (mois ture, temperature) (Mitchell & Parkinson, 1976;Luxton, 1981;Smrž, 1992band 1994. Reliable indications o f starvation in oribatid mites, starved for relatively short periods are changes in respira tion, dry body mass, glycogeneous granulae in mesen chymal tissue, gut contents and mortality.Significant changes in many microanatomical features and reproduc tive parameters only occur after prolonged periods (about 40 days) o f starvation.
mortality o f starved mites was significantly higher than o f the control mites after 21 and 42 days (log-linear analysis, x 2 = 3223, P < 0.001) and mortality was higher after 42 days than after 21 days (log-linear analysis x2 = 1909, P < 0.001, Tab. 1).Galumna elimata mites in the control group fed intensively on the green bark algae, Desmococcus vulgaris.Juveniles were observed after 42 days in the control group, but not in the starved group.Starvation significantly influenced the mass o f indi viduals after 21 days.The mean fresh mass o f fed indi viduals was 121 pg, dry mass 57 pg, and the water content 64 pg (53 % o f the fresh mass).The mean fresh mass o f starved mites was lower (113 pg).Their dry mass was 41 pg and the water content about 63 % (Fig. 1).Water content was significantly (P = 0.03) higher in the starved mites.The respiration rate was significantly lower in the starved than in fed mites.The respiration rate was 240 pl.g'fh'1 in the fed individuals and 130 pl.g'fh'1 (Fig. 2) after 21 days o f starvation.Starvation for 21 days did not influence the amylolytic activity.Mass specific amylolytic activity was 5 903 mg-1.g-1.h-1 and 5 762 mg-1.g-1.h-1 in the control and starved group, respectively (Fig. 3).The digestive tract o f Galumna elimata (Figs. 4, 5) is similar to the digestive tract o f the model species Ceratozetes cisalpinus (see Woodring & Cook, 1962b) and Euzetes globulus (see Hoebel-Mavers, 1967).TABLE 1. Mortality of starved and control Galumna elimata control Galumna elimata.DW dry mass, Water mass of body water, S starved group, C control group.The numbers in column = number of replicates.All parts of the gut o f the control mites contained algal food boli (Figs. 4, 5, 7).The guts o f the starved mites were full o f mucoid substances (Fig.
After 21 days o f starvation, indi viduals with mucoid boli prevailed over individuals lacking mucoid droplets in the mesenteron.Individuals lacking mucoid substances in the mesentron prevailed after 42 days.Generally, mesenteral cells were thicker in the anterior than in the posterior part o f the mesenteron o f the control

Fig
Fig. 3.A comparison after 21 days of the amylolytic activity in starved and control Galumna elimata.EA mass specific amylolytic activity, S starved group, C control group.
were present in the majority o f specimens from the control group and in some animals starved for 21 days, but apparently absent after 42 days of starvation.The salivary glands were lobed.Some cells contained a large nucleus and vacuoles.Other cells in the lobes formed reservoirs.There were two kinds (physiological types) o f salivary glands in control individuals; namely those without reservoirs with vacuolized cells (Fig. 19) and glands with well developed reservoirs.The second type prevailed in the starved mites (Fig 20).
. The mucoid bolus could be formed just before food ingestion, and the ingested food particles are then packed in this mucus (cf Hoebel-Mavers, 1967).Starvation led to changes in the micromorphology o f the reproductive tract and then reproduction is suppressed.The reduction in the size o f the seminal vesicle cells and the absence o f spermatic cells in the seminal vesicles is correlated with the high energy costs o f maintaining spermogenesis.Under natural conditions spermatic cells are present in a high proportion o f oribatid species throughout the year (Smrž, 1992a; Hubert & Smrž, 1998).Thus this feature is a suitable indicator o f starvation.The lower mortality o f gravid females may be because they contained more storage compounds at the beginning o f starvation.The origin o f the clumps o f cells in the oviducts o f some o f the starved females is unclear.