Z. Säugetierkunde61(1996)352-364 ZEITSCHRIFr^^^FUR © SAUGETIERKUNDE 1996GustavFischer,Jena INTERNATIONALJOURNAL OF MAMMALIAN BIOLOGY Size and structure ofburrow Systems ofthe fossorial rodent Ctenomys mendocinus in the piedmont ofMendoza province, Argentina By Maria I. Rosi, Mönica I. Cona, Silvia Puig, F. Videla, andV. G. Roig InstitutoArgentinodeInvestigacionesdeZonasAridas, UnidaddeZoologiayEcologiaAnimal, Consejo NacionaldeInvestigacionesCienüficasy Tecnolögicas, Mendoza,Argentina ReceiptofMs. 05. 07. 1995 AcceptanceofMs. Ol. 06. 1996 Abstract The structure and size ofburrowSystems of Ctenomysmendocinuswere analysed in the present study. These burrows showed a hnear pattern, with a main axis from which branches and laterals forked of BurrowSystemsofmaleswerelargerthanthoseoffemales,withalongermaintunnelandgreaternum- berofbranches. Lengthening ofthe main tunnel was achievedby addingnewsegmentsratherthan by excavating longersegments. Both male and female burrows showedthe same geometricconfiguration. There were no differencesin the percentagesestimatedformaintunnel,branches andlaterals. The an- gular variables (directional angle, brauch angle and angle of ascent of laterals) did not show any differencesbetweenboth sexes either. Burrows showed a constant heading along theirpathwith mean directional angles close to 0°, even though in most Systems left- and right-ward deviations from the main tunnel followed a random sequence. Branches originated at right angles to the main tunnel. C. mendocinus appears to minimize the energy cost ofburrowing by increasing the angle of ascent of the lateral instead ofits length as the maintunnelgrows deeper. Homeränge,perimeter, andlinearity weresignificantlyhigherinmales. Introduction Fossorial mammals, just as subterranean mammals, are largely confined to Underground life, but they venture a few centimeters outside their burrows for foraging (Contreras and McNab 1990). Both groups of herbivorous mammals construct complex tunnel Sys- tems, which ensure them a relatively constant microclimate, protection from predators and access to food (Nevo 1979; Reichman and Smith 1990). The subterranean environment is largely constant, and selection has led to the devel- opment of highly convergent structural features in the burrow Systems ofdiverse mamma- han taxa (Nevo 1979; Hickman 1990). Knowledge of burrow structure and its adaptive features is a major aspect in understanding the unique evolution and adaptive biology in subterranean mammals (Hickman 1990). The architecture of burrow Systems of the species of Ctenomys represents an almost unknown aspect ofthe ecology ofthis South American endemic fossorial rodent. Informa- tion available is either fragmentary or based on the study of only a few animals. Linear Systems composed of a main axis and lateral tunnels have been reported for C. opimus, C.peruanus (Pearson 1959) and C.pearsoni (Altuna 1983). A similar pattern was de- scribed for C. mendocinus in an environment of low productivity (Puig et al. 1992). Line- ar Systems associated with low resource availability have also been reported for Thomomys bottae (Reichman et al. 1982), Tachyoryctes splendens (Jarvis and Sale 1971) and Spalax ehrenbergi (Heth 1989). SizeandstructureofburrowsofCtenomysmendocinus 353 This study aims at elucidating the most relevant structural features of burrow Systems of C. mendocinus, and their possible differences between sexes. Material and methods Burrow Systems of C. mendocinus were studied at Las Higueras (Mendoza, Argentina, 32°30'S and 68°55'W), locatedinthefoothillsoftheAndes (1125melevation). Thezonepresentslowhillocksand depressions furrowed by wadis. Solls are made up oflithosols in a matrix of strongly compacted clay Fig.1. SchematicrepresentationsofexcavatedburrowSystemsofCtenomysmendocinusfromthepied- montofMendoza(Argentina).Burrowsofadult(A)maleand(B)female.Tu=maintunnel, Br=brauch,La=lateral. 354 Rosi,MariaI.,etal. andsilt. ShrubcommunitiesofLarreadivaricata and L. cimeifolia prevail, alternatingwithdenseripar- ianVegetation(Roig 1976). CompleteburrowSystemsof12 adults (6males,6females) were excavatedandmappedin Septem- ber and October 1991. Afteroccupants were captured, Oneida Victorkill-trapswere kept for24hours in the burrowopenings, in orderto recordthe numberofoccupants perSystem. Sex, weight, reproduc- tiveconditionandrelativeageweredetermined,basedoncriteriaofRosietal. (1992). Length,depth,diameteranddirectionalityoftunnelsectionsandChamberswere measuredforevery System. Burrow openings and surface mounds were mapped. Burrow structural features were described with the same terminology used for other subterranean mammals (Vleck 1981; Reichman et al. 1982; Andersen 1988; Hickman 1990). The main tunnel was defined as the longest axis ofthe System (Fig. 1). When a fork was found, the longest tunnel was selected to follow the path ofthe main tunnel (Reich- mman et al. 1982). Mostselectedtunnelsshowedthe smallestangulardeviationfromthemain axis (An- dersen 1988). Branching tunnels forking offofthe main tunnel were classified into: a)laterals: straight tunnels that never brauch, and end in a surface opening with or without a soil mound. Laterals also in- cludedall those tunnelsendingin a "cul-de-sac'\ hereafterreferredtoasblindlaterals,b)branches: tun- nelsmadeupofatleastonesegmentandonelateral.Suchtunnelssometimeshavesecondarybranches. InburrowSystemsdescribedforotherfossorialrodentsthe term"segment"isusedtodefinetunnel sections that extend between laterals (Vleck 1981; Reichman et al. 1982; Andersen 1988). In C. mendociniis, these tunnel sections were comparatively longer and showed great variabihty, owing to the small number of laterals recorded. For this reason, every fragment of such tunnel sections deter- mined by a change in heading was called segment, therefore our segment is not equivalent to the one definedbytheabove authors. Oneofthe endsofthemaintunnelwasarbitrarilychosenastheinitialsegmentoftheSystem,since we did not know the sequence ofsegment construction. Deviation angles (directional angles) between consecutive segments of the main tunnel were measured from that initial segment, considering their left-ward (negative) orright-ward (positive) deviation. The mean ofdirectional angles permitted us to knowSystemdirectionality.accordingtoAndersen (1988).Thebrauchanglewasmeasuredasthesmal- lest angle betweenthe initialsegmentofeverybrauchandthe maintunnel.The angle ofascent (Vleck 1981)wasestimatedonlyforthoselateralsthatreachedthesurface (bhndlateralswereexcluded). TheareaandperimeterineverySystemweremeasuredwithTecktronix4958digitizerusingthe IN- CYTH-CRAs Programme for calculating areas. For this purpose, an irregulär polygon was drawn, by joining the ends of all laterals in a clockwise direction. For every angle between two segments from which no lateral came off, the bisector ofthe convex angle was drawn. The length ofan estimated lat- eral was projected on this bisector. This lateral was obtained by averaging the perpendicular distances between the end ofevery actual lateral and the originating tunnel segment. Linearity ofa burrowSys- temwasdeterminedbasedoncriteriaofReichmanetal. (1982). AngularvariableswereanalysedbyZar's(1984)circularstatistics,usingtraditionalstatisticsforthe remainingvariables. Allvariableswerefirsttestedforhomogeneityofvarianceusingthevarianceratio test (Zar 1984). Accordingtoresults,theStudent'st-testorMann-WhitneyUfestwereusedtoanalyse fordifferencesbetweensexes. An approximate testingprocedure fordifferences between twoproportionswas usedas appliedby Zar (1984). Linear regressions were performed between different metric variables estimating data ad- justmentbyusingPearson'sproduct-momentcorrelationcoefficientr. Results Size andstructureofsubterranean Systems All excavated Systems (n= 12) corresponded to sexually active adult animals. No signifi- cant differences were recorded (t = 1.69; P<0.10)_in dry lens weight between males (n = 5; X = 322mg; SD = 2.9) and females (n = 6; X = 29.0mg; SD = 3.19). In contrast, males (n = 5; X =219_g; SD = 52.7) proved to be significantly heavier (U = 27; P = 0.025) than females (n = 6; X = 140.5 g; SD = 10.8). Wide ranges of body weigth were obtained within each sex (140 to 260g for males, and 126 to 155 g for females). Every burrow was inhabited by only one specimen, except for a female's burrow with five youngs in the nest. SizeandstructureofburrowsofCtenomysmendocinus 355 Total length of tunnels in a System ranged from 15 to 31 m for females, and from 23 to 77m fqr males. Significant differences were found (U = 34; P = 0.005) between male (n = 6; X = 50.5 m; SD = 21.0) and female burrows (n= 6; X =22.4m; SD = 6.8) concern- ing the average of the total length of tunnels. The diameter of feeding tunnels was largely constant, with mean values of 8.03 cm (n = 17; SD = 1.72) for females, and 8.5 cm (n = 33; SD = 1.51) for males. Slight enlargements were only observed in the branching points of the main tunnel. Main tunnel: The main tunnel represented the greatest proportion of the total length of the System (males: 63.9%; females: 60.3%), there being no differences between sexes (Z= 0.68; P<0.50). Total length and number of segments ofthe main tunnel were signifi- cantly greater in males (Tab. 1). Segment length varied within a wider ränge for males (0.20 to 3.20m) than for females (0.20 to 1.60m). Mean segment lengths were not signifi- cantly different either within or between sexes (Tab. 1). The main tunnel length showed a significant correlation with the number of segments (r= 0.92; df= 10; P<0.001), with their mean length (r =0.78; df=10; P<0.005), and with the occupant's body weight (r=0.69; df= 10; P<0.02). In spite ofthe width ofweight ranges within each sex, correla- tions with the main tunnel length were not significant. The main tunnel depth showed little Variation both within and between burrows; moreover, no differences were found related to sex (Tab. 1). Greatest depths recorded did not exceed 0.40m. The directional angles of consecutive segments of the main tunnel ranged from -128° to 132° (n = 350). Figure2 shows the bimodal and symmetrical distribution of directional Table1. Comparisonbetweensexesofmetricvariables(X±SD)inburrowSystemsofC. mendocinus, wheretandUarethestatisticsoftheStudent'slestandMann-Whitneytest,respectively;nisthenum- berofvaluesand(k)thenumberofmeansusedtoestimatesecond-ordermeans. Variable Females Males Valueofthe Significance n (k) X+SD n (k) X±SD statistic level MAINTUNNEL Totallength(m) 6 13.53±4.42 6 32.31±14.05 U=33.00 p=O.Ol Numberofsegments 6 26.50±9.01 5 41.60±11.91 t= 2.39 p<0.025 Meansegmentlength(m) 159 (6) 0.52±0.26 208 (5) 0.68±0.46 t = 0.78 p<0.25 Meandepth(m) 112 (6) 0.30±0.09 227 (6) 0.27±0.06 t= 0.63 p<0.25 BRANCHES Numberofbranches 6 2.50±1.22 6 4.00±1.51 t = 1.86 p<0.05 Meanlength(m) 15 (6) 1.54±1.07 24 (6) 2.25±0.63 t= 1.40 p<0.10 Meannumberofseg- ments 15 (6) 2.96±1.46 24 (6) 3.06±0.63 u= 18.50 p>0.10 Meansegmentlength(m) 15 (6) 0.47±0.26 24 (6) 0.72±0.61 t= 0.92 p<0.25 LATERALS Meannumberoflaterals 6 15.83±4.22 6 22.83±10.26 t= 1.55 p<0.10 Meanlength(m) 95 (6) 0.34±0.16 137 (6) 0.40±0.24 t= 0.49 p>0.25 Meandepth(m) 86 (6) 0.26±0.08 136 (6) 0.26±0.06 HOMERANGE Area(m^) 6 11.91±6.08 6 43.13±31.13 u=33.00 p=O.Ol Perimeter(m) 6 31.67±11.00 6 74.43±33.33 u=34.00 p=0.005 Linearity 6 2.60±0.36 6 3.27±0.57 t= 2.41 p<0.05 356 Rosi,MariaI.,etal. angles > and <0° in all twelve Systems. Mean values for each System varied between -11° and 8° (n = 12) (Tab. 2), showing no significant differences either within or between sexes (Tab. 3). Moreover, when considering separately the positive and negative directional an- 60 I 1 Directional Angle (degrees) Fig.2. Frequencydistributionofdirectionalangles(n=350)inmaintunnelsofburrowSystemsof C. mendocinus. Table2. Meanvalues (degree) andangulardeviations(S) ofthedirectionalangleformaintunneland branches. Animal Meandirectional Directionalanglesofmaintunnel(*) Meandirectional anglesofmain anglesofbranch tunnel (+and-) Positive Negative segment(**) Num. Sex n X S n X S n X S n X S 214 15 8 40 8 39 30 5 -36 20 215 ? 14 -3 43 7 40 23 7 -42 10 228 ? 34 -1 49 16 47 23 16 -50 24 7 9 65 269 ? 26 -11 50 11 44 25 14 -53 25 4 16 27 270 ? 31 1 40 16 33 23 14 -37 24 11 18 43 272 ? 32 4 43 14 47 25 14 -36 23 4 14 72 216 22 -5 52 11 45 25 11 -54 23 4 -80 56 218 57 -1 52 29 45 26 26 -56 30 7 -19 46 224 27 —3 48 14 39 34 11 -52 17 12 -1 43 225 39 6 56 20 58 30 17 -53 25 16 0 45 227 s 48 1 43 18 48 22 17 -49 24 4 -5 44 229 36 -1 46 17 42 30 17 -41 22 ('^) Accordingtoleft-ward(negative) andright-ward(positive)deviationsfromthemaintunnel. (**) EstimatedonlyforSystemswithbranchescomposedofmorethantwosegments. SizeandstructureofburrowsofCtenomysmendocinus 357 gles in every System, differences in number and in mean values were found to be minor (Tab. 2). By use of the one-sample test for the mean angles (Zar 1984) it was verified for every burrow that the mean directional angle did not deviate significantly from 0°. No significant deviation was recorded when analysing the mean values obtained for each sex (Tab. 3) at the 99% confidence level (length of mean vector r= 0.62 in males and 0.70 in females). Table3. Comparisonswithineachsex(usingChi-squarecontingencytest) andbetweensexes(using Watson's test)ofangularvariables(X=mean,S=angulardeviations)ofburrowSystemsof C. mendocinus,wherenisthenumberofvalues,(k)thenumberofmeansusedtoestimatesecond- ordermeansandPlevelofsignificance. Sex Directionalangle Brauchangle Angleofascent ofmaintunnel oflateral Females n(k) 152(6) 13 31(6) X±S 0.1 ±44.7 88.1±18.1 40.4±18.3 Chi-square(df) 45.6(35) 23.4(18) 44.3(35) p<0.25 p<0.25 p<0.25 Males n(k) 229(6) 21 39(5) X±S 0.4±49.8 85.8±19.9 31.1±35.4 Chi-square(df) 46.4(45) 30.0(20) 25.6(24) p<0.50 p<0.10 p<0.50 Comparisonbetweensexes U'(df) 0.095(6.6) 0.068(13,21) 0.103(6,5) p<0.20 p<0.50 p<0.50 To test whether the sequence of construction of right-ward and left-ward segments was random, the two-tailed runs test (Zar 1984) was used. In only three ofthe twelve Sys- tems considered was the null hypothesis for randomness rejected, which indicates that in most Systems segment deviations towards either side ofthe main tunnel did not alternate. Brauches: in every System, one to six branches forked off from the main tunnel. The mean number of branches per System was significantly greater in males than in females (Tab. 1), whereas the mean number of branches per meter of main tunnel (males = 0.13; SD = 0.03; females = 0.20; SD = 0.14) did not differ significantly between sexes (U = 25; P>0.10). Brauch mean length, as well as length and mean number of brauch segments, showed no variations related to sex (Tab. 1). The proportion of branching tunnels in the total System length did not _differ significantly (Z= 0.32; P>0.5) between male (X = 17.5%) and female burrows (X= 16.2%). Brauch number, but not mean length, was significantly correlated with the main tunnel length (r=0.79; df= 11; P<0.002). The highest frequencies ofbrauch angles ranged between 80° and 100° (Fig. 3). Means of these angles were similar for both sexes (Tab. 3) and did not depart significantly from 90° (length ofmean vector r= 0.95 in females and 0.94 in males, one-sample test for mean angles, at 99% confidence level, Zar 1984). Most branches composed ofmore than two segments showed a mean directional angle close to 0° (Tab. 2). Brauch depth (X = 0.26m; SD = a06) did not differ significantly (t = 0.75; df= 22; P<0.25) from the main tunnel depth (X =0.28m; SD = 0.07). 358 Rosi,MariaI.,etal. 8 I 65 85 105 125 Branch Angle (degrees) Fig.3. Histogramofbranchangles(n=34)measuredinallexcavatedSystemsofC. mendocinus. Laterals: 8 to 37 laterals per System were found. Their mean number was significantly lower than the mean number of segments composing the System, both in males (t = 3.8; df= 10; P<0.001) and in females (U = 32; df= 6.6; P<0.025). No significant differences between sexes were recorded in the mean number of laterals (Tab. 1) or in the proportion of laterals relative to total segments (females: 47%; males = 45%; Z = 0.45; P> 0.50). The length of main tunnel sections, comprised_between two laterals reaching the surface, showed a wide ränge of Variation (n = 71; X= 3.7 m; SD^3.9). This high variability was also observed when blind laterals were included (n = 151; X = 1.8m; SD = 1.8). Lateral length varied considerably both within and between burrows (0.15 to 1.50m), m although few laterals exceeded 0.90 in male (6.6%) and female (1%) Systems. Lateral depth at the originating point was not significantly correlated with lateral length. Nevertheless, depth and angle of ascent of laterals were correlated (r= 0.46; df= 69; P< 0.001). These three variables showed no significant differences between sexes (Tabs. 1 and 3). Angles of ascent varied between 21° and 58° with their highest frequen- cies between 15° and 45°, and only 7% were higher than 60° (Fig. 4). Total length of laterals in proportion to total length of System tunnels was similar for both sexes (males: 19%; females: 23%; Z=1.10; P<0.50). The proportion of laterals reaching the surface (54.7%, n = 127) was similar to that of blind laterals (45.3%). The latter were usually as deep as the main tunnel; 11.4% of them (n = 105) were plugged with shredded plant material mixed with loose soil. About 25% of laterals (30.4% of their overall length) were totally or partially plugged with slightly compacted soil. Chambers: eight Systems showed oval Chambers (1 to 3) that outsized the diameter of feeding tunnels. Considering their contents and location. they were classified into nest Chambers, and feeding-resting Chambers. No defecation Chambers were found. SizeandstructureofburrowsofCtenomysmendocinus 359 35 55 Angle of Ascent (degree) Fig.4. Anglesofascentoflaterals(n=70)measuredinelevenburrowSystemsofC. mendocinus. The Chamber located at a greater depth than the remainder of the System with dry plant material covering its walls was regarded as the nest. Nests occupied an eccentric Po- sition in most Systems, never being situated in the distal ends ofmain tunnels or branches. No significant differences were recorded (t = 0.64; V>0.25) in nest size between males (n=4; X= 0.32 m;_SD = 0.005) and females (n =4; X=0.29m; SD =0.003). Depth of fe- male nests (n = 4; X= 0.64_m; SD = 0.04) was significantly greater (t = 3.16; P<O.Ol) than that ofmale nests (n = 4; X = 0.34m; SD = 0.001). Feeding-resting Chambers were smaller than nests and were located in the main tun- nel, at the same depth. They were usually empty, although fresh plant material was found in some of them. Storage ofplant material was found into little blind laterals in most bur- rows. Mounds: the mean number of mounds recorded per burrow was 6.2 (n= 10; SD = 3.08). Most of them were roughly circular, with a mean diameter of 0.52 cm (n=21; SD = 0.14). Connection of these mounds with the burrow could not be found in 44% of them, in spite oftheir adjacent position to the main tunnel or blind laterals. Nearly all the mounds (95%) were old, given their flattened shape and high degree ofsoil compactness. Home ränge Area, perimeter and linearity were significantly higher in male than in female burrows (Tab. 1). Home ränge size varied between 7 and 16m^ in females, and between 14 and 99m^ in males. In both sexes, the perimeter of home ranges (n = 12; X = 53.1 m; SD = 32.5) departed significantly (U = 135; df= 12.1_2; P<0.005) from the value expected for a circular System having the same area (n = 12; X = 17.0m; SD = 7.8). 360 Rosi,MariaI.,etal. Area (r= 0.92; df= 11; P<0.001), perimeter (r = 0.98; df= 11; P<0.001), and linearity (r= 0.69; df=ll; P<0.01) were significantly related to main tunnel length, but not to mean branch length. However, both area (r = 0.65; df=11; P<0.02) and perimeter (r = 0.74; df=ll; P<0.005) proved to be significantly correlated with the number of branches in every System. Discussion BurrowSystem size C. mendocinus males constructed larger Systems than females with a longer main tunnel and a greater number of branches. Differences between sexes in the main tunnel length were caused by a greater number of segments, rather than by longer segments. Sexual dimorphism in body weight and burrow length coincide with records for T. bottae (Reichman et al. 1982), and S. ehrenbergi (Heth 1989). For this latter species, Heth (1989) suggested that burrow length may be related to food requirements rather than to sex and reproduction, based on the significant correlation between main tunnel length and body weight in a joint analysis of both sexes. For C. mendocinus this correla- tion was significant when sexes were pooled but not when each sex was considered sepa- rately, in spite of the width of weight ranges in either sex. These results suggest that sex rather than body weight affected the length of burrow Systems which, in males, can be re- lated to the need to contact potential mates during the reproductive season. The basic building units determined for Systems of T. bottae (Vleck 1981; Reichman et al. 1982) and Geomys biirsarius (Andersen 1988) could not be identified for C mendocinus Systems, owing to the scarce number of laterals. Increased compactness of the soil plugging the laterals could cause loss of laterals. There are several points of evi- dence that strengthen this assumption: a) great number of laterals plugged with slightly compacted soil. b) high percentage of soil mounds near the System but showing no evi- dent connection with it, and c) presence of numerous bhnd laterals. The latter might have originated from increased compactness of partially plugged laterals, although a small per- centage ofthem may reflect probings for food as suggested by Hickman (1983). Tunnel depth andslope oflaterals Feeding tunnels and Chambers composing the burrow Systems of C mendocinus were deeper than expected. considering the high energy cost of burrowing that a fossorial ani- mal must afford in cohesive soils (Vleck 1979, 1981). Although deeper Systems demand higher energetic expense (Vleck 1981), they are more stable regarding temperature and humidity (Llanos 1947; Rosenmann 1959; Altuna 1985). This stability is a factor of great importance in arid regions, where free access to water sources is usually restricted (Hickman 1990). Several authors have reported the importance of the level of subterranean parts of plants for determining the depth of feeding tunnels of subterranean rodents (Miller 1957; Pearson 1959: Jarvis and Sale 1971; Heth 1989: Williams and Cameron 1990). This did not seem to be the most important factor accounting for System depth in C. mendocinus, since the record of foraging signs above ground (stems cut on the bias near burrow openings) indicates that feeding is not based exclusively on subterranean parts ofplants. C. mendocinus seems to minimize burrowing cost by increasing the angle of ascent of laterals rather than their length, as the main tunnel deepens. However, values over 60° are not very frequent, which would counteract the tendency of loose soil to fall back into the burrow, thus being helpful in plugging the laterals. SizeandstructureofburrowsofCtenomysmendocinus 361 Characteristicsofmounds The flat shape, great size and compacted soil of mounds reveal that their construction was not recent, which indicates little burrowing activity during the study period (Septem- ber and October) that coincides with parturition and lactation. Throughout this season, fe- males invested more energy in htter care than in burrow expansion, whereas males completed expansion of their Systems searching for mates during the pre-reproductive season (winter). These behaviours have been reported for other species of subterranean rodents (Hansen 1960; Miller and Bond 1960; Bandoli 1981; Sparks and Andersen 1988; Busch et al. 1989; Rosi et al. 1992). On the other hand, backfilhng ofold tunnels ac- counting for the temporal Separation between excavation and mound-production rates ob- served in G. bursarius (Sparks and Andersen 1988; Andersen 1987; Thorne and Andersen 1990) was not evident in C. mendocinus given the low percentage of backfilled tunnels (excluding laterals). The small proportion of mounds found in Systems of C mendocinus, related to the number of laterals reaching the surface, partly could be due to the use of loose soil for plugging laterals. However, a high percentage of mounds might disappear as a conse- quence of rainfall runoff. Violent summer storms are usual in the piedmont area, giving origin to torrential streams capable of moving considerable amounts of Sediments (Roig 1976). Structural complexityofburrowSystems In general, burrow Systems of C. mendocinus follow the geometry described for other soli- tary fossorial rodents such as T. bottae (Vleck 1981; Reichman et al. 1982), G. bursarius (Andersen 1988), C.pearsoni (Altuna 1983), C. opimus and C.peruanus (Pearson 1959). In all these species burrows show a linear configuration, with a main axis com- posed of a sequence of straight short segments, and a variable number of branches and laterals forking off along such an axis. Burrows of C. mendocinus show a similar geometric configuration in both sexes. Not only the proportions of main tunnel, branches and laterals per System were similar be- tween sexes but also all angles considered. Mean directional angles close to 0° and similarities in number and mean value of left and right deviations from the main tunnel proved a constant heading along the Systems' path. Reichman et al. (1982) believe that the linear configuration of burrow Systems in T. bottae enables males to intercept a greater number of female burrows. Andersen (1988) concluded that the directed movement used by G. bursarius to locate resources would also be optimal for contacting females. Branches are originated perpendicularly to the main tunnel and tend to keep the right angle along their path. Andersen (1988) reported that both linearity and orthogonal branching of G. bursarius Systems were consistent with the search path predicted for a "harvesting animal" (Pyke 1978) from the optimal-foraging theory. Nonetheless, Ander- sen (1988) does not discard the influence of physical and physiological constraints on the burrowing process as being responsible for the geometric pattern. Home ränge C. mendocinus is a solitary species, even during the reproductive season (Puig et al. 1992). Moreover, absence of linking tunnels between burrow Systems suggests a strongly territorial behaviour for this species. In subterranean rodents, home ränge is considered to be coincident with territory, and is restricted to the burrow System (Ingles 1952; Howard and Childs 1959; Miller 1964; Nevo 1979).