Temporal and spatial patterns of foraging in two syntopic ant species, Manica yessensis and Formica lemani, on Mount Fuji, Japan

Spatiotemporal characteristics of terrestrial foraging were studied in two ant species, Manica yessensis and Formica lemani, in a volcanic desert on the southeast slope of Mount Fuji, Gotenba, Japan. Both ants are common in this habitat, and they construct underground nests in this dry area with sparse vegetation. Nests of M. yessensis have multiple nest‐openings on the surface, whereas nests of F. lemani have very few openings, but their nesting and foraging areas overlap completely. A “mark‐and‐observe” method applied to M. yessensis demonstrated that worker ants of this species move between openings more than 3 m away. A study plot (6 m × 12 m quadrat) was set up, in which all nest‐openings of both species were mapped. Day‐long observations on numbers of foragers in this plot revealed that foraging M. yessensis are active in morning and evening, while F. lemani continues foraging all day, but both species cease activity at night. Associations between locations of foragers and nest‐openings differed significantly between the two species, that is, surface foraging of M. yessensis workers was largely confined to the vicinity of their nest‐openings, whereas foragers of F. lemani travelled far from their nest‐openings. The function of multiple nest‐openings in M. yessensis is discussed.


INTRODUCTION
In ecological studies of social insects, the number of queens (fertile gynes) in a colony is one of the most important characteristics.In ants, there are many species in which multiple fertile queens coexist in a colony, a phenomenon called polygyny, as distinguished from monogyny (Hölldobler & Wilson 1990).In many polygynous species, colony members are distributed in multiple, discrete or discontinuous nest sites, a phenomenon called polydomy, contrasting with the usual monodomy (Sundström 2021).In polydomy, each queen or queen group may have its own nest, and this condition may serve to split such branch nests to form new, autonomous colonies.However, polydomy is observed also in monogynous (single-queened) species.When the number and capacities of nest cavities are limited, as in the case of arboreal ants that inhabit dead branches or inside trees, the colony needs polydomy to harbor its large population (Torossian 1968;Hasegawa 1992).
However, even in the absence of physical limitations for nest sites, some monogynous ant species adopt polydomy (Cerd a et al. 2002).One species that does this is Manica yessensis Azuma, 1955, which occurs in montane areas of Japan (Terayama et al. 2014).It is one of the most common ants in the volcanic desert on the east and southeast hillsides of Mount Fuji.In this semi-arid habitat with sparse vegetation, M. yessensis colonies construct underground nests with many nestexits that open to the surface over several square meters.Therefore, a simple but important question is why so many nest-openings are necessary for this monogynous ant, because it appears that there is no limitation of space for construction of underground nests.In order to understand how these multiple nestopenings are used by M. yessensis, the present study investigated the distribution of nest-openings on the surface and movements of ants between nest-openings.Also, spatiotemporal patterns of foraging by worker ants, especially, in relation to nest-opening location was investigated in a study plot (6 m Â 12 m).
In the same habitat, another ant, Formica lemani Bondroit, 1917, is equally common, but this species is monodomous.Therefore, the foraging characteristics of F. lemani were also investigated and compared with those of M. yessensis.

Study site
The study site was located in the vicinity of the Gotenba Trail New 5th Station on the southeast slope of Mount Fuji (35.3362 N,138.7857E, alt.1,550-1,700 m), Shizuoka Prefecture, Japan.The area was previously used as a ski slope operated by Gotenba City, so that 17 steel poles of a ski lift stand across the slope over a distance of 690 m (Fig. 1a).As these poles can be used as landmarks to map ant nests, mapping of all nests of M. yessensis was conducted on this slope, in an area approximately 100 m wide and 700 m long.The surface of the east and southeast slope of Mount Fuji, including this 7 ha study site, is covered with a thick layer of scoria (volcanic ash and lapilli).At this site, vegetation is sparse, chiefly because of the mobility and dryness of the substrate (Ohsawa et al. 1971).This type of habitat is generally called a volcanic desert.However, there are still many patches of grasses and shrubs, and these plants nourish many insects.Therefore, nutrition of ants living in this habitat, feeding directly on photosynthates or indirectly on invertebrates, depends on photosynthesis of these plants (Ohsawa et al. 1971).

Ants
Manica yessensis and F. lemani are very common at this site.Manica is a small ant genus in the subfamily Myrmicinae, containing one species each in Japan and Europe and four species in North America (Bolton 2022).Recent molecular phylogenetic analyses confirm that Manica and Myrmica are sister groups (Ward et al. 2015).In contrast, Formica is a rather large genus of the subfamily Formicinae, containing 179 species (Bolton 2022) and widely distributed in the Holarctic region.Nine species of Formica are known from Japan (Terayama et al. 2014).Both M. yessensis and F. lemani are montane species in central Japan (Terayama et al. 2014).On Mount Fuji, the altitudinal distribution of M. yessensis is from 1,450 m to 2,380 m and that of F. lemani is from 1,450 m to 3,000 m (Kondoh 2023).Their workers are similar in size (body length 3.5-5.5 mm) (Terayama et al. 2014).

Polydomy, a multiple-nest system
Manica yessensis is a monogynous species (K.Masuko, unpubl. data, 1991).Its nests are excavated in soil, and a developed colony has many nest-openings on the surface.Nest-openings of M. yessensis can be easily distinguished because of their conspicuous crater shape (Fig. 1c).Such openings are distributed over a fairly large area (Fig. 2a,b), indicating that their nests are polydomous, because "they [social insect colonies] simultaneously occupy multiple, spatially separated nest sites" (Sundström 2021), which may or may not be structurally connected by tunnels.As there have been some arguments for definitions of "polydomy" or a similar term "polycaly" (Debout et al. 2007;Sundström 2021), the term "multiple-nest system" is used in this manuscript.In addition, an isolated aggregation of nestopenings is referred to herein as a "nest-area," because social connections between such separate nest-openings have not been completely confirmed by behavioral or genetic evidence, although multiple nest-openings in M. yessensis are clumped and separated from other such clumps (Fig. 2, inset maps).
In this 7 ha study site, 124 nest-areas of M. yessensis, as defined above, were found from 1988 to 1993.They were numbered, mapped and marked with stakes.In contrast, F. lemani is polygynous and constructs monodomous nests, which normally extend into the soil vertically along roots of knotweed, Fallopia japonica (Houtt.)Ronse Decr.(M.Kondoh, unpubl. data, 1975).These nests may have multiple nest-openings on the surface, but their spread is limited to a smaller area (less than 1 m 2 ), and at the depth of 0.3-0.5 m, the galleries from these nest-openings unite into a single vertical shaft that extends deeper (M.Kondoh, unpubl. data, 1975).In the present study, no survey was conducted on the distribution of F. lemani nests in the 7 ha area.
In order to map nest-entrances of the two species, a compass (for angles) and a tape measure (for distances) were used.This measurement started from the base stake, the location of which was determined from the nearest of the 17 ski lift poles.To examine whether Manica ants stayed below each opening, the soil below and around openings was excavated with a hand trowel.

Mark and observe
As multiple nest-openings of M. yessensis were clumped and separated from other such clumps, it seems that a cluster of nest-openings belongs to a single colony.To confirm this, each opening was marked with a numbered chopstick flag, and workers going in and out of openings were collected with an aspirator and immediately marked on the gaster with a colored spot of oil paint (Uni Paint Marker PX 21; Mitsubishi Pencil Co., Ltd.).Different colors were used for different openings.When too few ants were collected on the surface, part of the nest-opening was excavated by a hand trowel to collect ants.Marked ants were released at that opening immediately after being marked.Searching for marked ants was conducted at a later date.When marked ants were rediscovered, few were collected.Therefore, the numbers of sightings, not numbers of collected ants, were analyzed in this investigation.This markand-observe study was carried out on three nest-areas, #88-1, #88-57 and #88-72.A total of 201 workers were marked for #88-1, 425 workers for #88-57 and 330 workers for #88-72.However, for the latter two, observations were too few.At #88-57, there were 12 observations of marked ants, of which five were at their marked openings and seven at different openings.At #88-72, there were six observations of marked ants, of which two were at their marked openings and four at different openings.Therefore, only the results of #88-1 are reported here.Its workers were marked at five openings on 1 July 1994.Ants from different openings were marked in different colors, that is, 70 with light blue, 54 with pink, 11 with yellow, 4 with red and 62 with white.After marking, a census was conducted on 10, 15, 24 and 29 July 1994.

Diurnal rhythm of foraging and spatial foraging patterns in relation to nest-openings
A quadrat plot (6 m Â 12 m) subdivided into 1 m squares was laid out to investigate temporal and spatial patterns of foraging by M. yessensis, and concurrently, to compare with syntopic F. lemani.The plot was positioned in a more or less flat place (Fig. 1b), where large numbers of nest-openings of both species existed.The plot was divided into 72 quadrats, each 1 m 2 , with a numbered chopstick flag at each intersection.Locations of nest-openings of both ant species were then mapped in each of these 72 quadrats.Mapping of M. yessensis nest-openings started on 18 June 1993 and that of F. lemani on 9 July 1993.Totals of 42 and 31 nest-openings were mapped for them, respectively, by the time the first census of surface ants started on 28 August 1993.One or two additional nest-openings were found right before or during the second census, carried out on 12 September 1993, when total numbers of mapped nest-openings reached 43 and 33, respectively.On that day, positions of stems or trunks of plants in the study plot were recorded, except for those growing alone, most of which were Aconogonon weyrichii (Polygonaceae) (Fig. 3).At each census, numbers of Manica and Formica workers in each quadrat were counted and all quadrats were traversed in a fixed order.Only foraging ants were counted, and not ants engaged in nest maintenance.The latter workers were easily distinguished because they remained at or went in and out of nest-openings.
On the surface of the study plot, workers of other ant species, Lasius japonicus Santschi, 1941, Formica yessensis Wheeler, 1913 and unidentified (perhaps two) species of Myrmica, were also observed during the present study.However, because quadrats where these ants appeared were limited, and their numbers were few, it appeared that their presence had little or no effect on foraging of M. yessensis and F. lemani.Numbers and locations of these other species were also recorded, but not considered in subsequent analysis.In the study plot, four small, isolated island-like plant communities existed, which included Larix kaempferi (Lamb.)Carrière (saplings), Salix reinii Franch.& Savat.ex Seemen, Spiraea japonica L.f., Aconogonon weyrichii var.alpinum (Maxim.)H. Hara, Astilbe thunbergii (Siebold & Zucc.)Miq.var.fujisanensis (Nakai) Ohwi, Fallopia japonica and species of Poaceae.Locations of these plants were also mapped.
The first survey was undertaken on 28-29 August 1993, in which ants were counted every 3 h from 18:00 h until 18:00 h the next day.As this survey showed that both species ceased foraging activity at night, the second survey was conducted every 2 h from 03:00 h to 19:00 h on 12 September 1993.A flashlight was used in the dark and each census took approximately 1 h.The advantage of this method was that both temporal and spatial patterns of ant surface activities could be investigated concurrently.In both surveys, weather during the daytime was sometimes sunny and sometimes cloudy (covered in mist), but there was no rain.Mount Fuji is an independent mountain, and in summer time, moisture rises from the foot every afternoon and a large area, including the study site, was often hidden in mist.

Statistical analyses
Statistical tests were carried out using Statistix 10 (Analytical Software).

Multiple nests and numbers of residents
Two Manica nest-areas (#88-47 and #88-74) were investigated for presence of adults and immature broods beneath surface openings (Fig. 2a,b).For nestarea #88-74, 32 openings were investigated on 30 August 1990.These openings appeared to be separated into two clumps (Fig. 2a, upper-left and lower-right groups).No immature brood was excavated under any openings of the upper-left group, but there were only a few workers at most.In the lowerright group, there were three sites where immatures and/or alates (males and gynes) were found.Therefore, it is likely that both groups belong to the same colony and that the upper-left group was a branch, whereas the lower-right group was the main colony with brood chambers.Several nest-openings of F. lemani were found adjacent to and between Manica nest-openings (Fig. 2a).
Similarly, 21 openings of nest-area #88-47 were investigated on 30 August 1990 (Fig. 2b).At four single-opening sites or a few adjacent openings, larvae or pupae, or larvae with alates were found in addition to workers.At each of the other openings, 0-10 workers were found.

Mark and observe
Nest-openings of #88-1 appeared to be separated into two areas (Fig. 4).During censuses, 29 previously marked workers were observed, and 18 of those were found at openings different from those where they were marked (Fig. 4).In addition, on 24 July 1994, at the base of a thistle (Cirsium purpuratum) 1.1 m from the nest opening where ants were marked with white (Fig. 4, Th), a few Manica workers were excavating.There were also workers coming and going between that opening and the thistle (Fig. 4, asterisk), and one worker with a light-blue mark was found under leaves of the thistle.They were likely constructing a new nest entrance.Ants were paint-marked with different colors at five nestopenings (solid circles) on 1 July 1994; numerals following the color are the number of marked ants.Dashed arrows indicate where and how many marked ants were observed in later censuses.Several marked ants were found in brood chambers (BC) that had no surface openings.A few M. yessensis workers were observed walking between a nest-opening and a thistle (Cirsium purpuratum) (asterisk and Th) and excavating at the base of the plant (observation on 24 July 1994).
Therefore, movements between two openings even more than 3 m apart were confirmed.These movements of marked ants demonstrated that all openings depicted in Figure 4 belong to the same colony.

Diurnal rhythm of foraging
Locations of nest-openings of M. yessensis and F. lemani in the 72 m 2 study plot are shown in Figure 3. Numbers of ants counted from all 72 quadrats at each census were plotted against census times (Fig. 5).The two studies on different dates show similar temporal characteristics of each species.Foraging of M. yessensis was crepuscular (Fig. 5a,c), that is, its activity showed two peaks, one in the morning and the other in the evening, while it remained inactive on the surface at midday.This ant completely stopped surface activity at night.On the other hand, foraging of F. lemani was continuously diurnal (Fig. 5b,d), but it also stopped almost completely at night.

Spatial patterns of foraging
Spatial distributions of foragers were investigated with reference to locations of nest-openings of the respective species.In each of the 72 quadrats, the association between nest-openings and foraging ants was examined (Fig. 6).Numbers of ants recorded in each quadrat were accumulated over all censuses (nine censuses for 24 h in the first survey and nine censuses for 16 h in the second survey), and total numbers were used as frequencies of foragers in each quadrat.Associations were examined by χ 2 -tests (Table 1).With M. yessensis, few foragers were observed in quadrats that lacked nestopenings (Fig. 6a).On the other hand, foragers of F.lemani were seen in almost all quadrats, irrespective of the presence or absence of its nest openings (Fig. 6b).It seems that, unlike M. yessensis, foragers of F. lemani forage at greater distances from their nests.
In total, 117 M. yessensis workers were recorded from all 72 quadrats, of which only one was on a plant.In contrast, 637 F. lemani workers were observed, of which 137 were found on plants.Therefore, densities of foragers in the same study plot differed greatly between the two species.Some correction may be needed in order to make a comparison, and one plausible solution was to adjust densities of both species to the same magnitude.Then, the total number in F. lemani was adjusted from 637 to 117.Accordingly, the frequency of F. lemani workers in each quadrat was reduced proportionately, and numbers were rounded to one decimal place.These corrections changed the frequencies of many quadrats to zero (Fig. 6c), so the effect of this correction was conservative.In the first survey conducted in August, there were 44 quadrats that had no nest-openings of M. yessensis, and Manica foragers were witnessed in only 16 of them (Table 1).In contrast, there were 47 quadrats that had no nest-openings of F. lemani, but foragers of this species were witnessed in most of them (44 of 47).The difference in these ratios was highly significant (χ 2 = 30.6651,d.f.= 1, P < 0.00001).Even after correction, the difference was highly significant (χ 2 = 6.8287, d.f.= 1, P = 0.00897).Very similar results were also obtained from the second survey, conducted in September (Table 1).In this survey, the total number of Manica workers recorded from all 72 quadrats was 91 (all on the ground), whereas 427 F. lemani (301 on the ground and 126 on plants) were observed.Therefore, surface foraging of Manica workers was much more confined to the vicinity of nest openings, compared to F. lemani, foragers of which travel far from their nests.

Agonistic interactions between M. Yessensis and F. Lemani
Surface foraging areas of M. yessensis and F. lemani overlap completely, so there could be substantial competition for food and possibly also for nest sites.However, it was rare to witness agonistic interactions at this study site.Only two instances of strife were observed.The first was on 23 July 1988 (around 15:00 h) when, at a nest-opening of F. lemani, which was near the nest-area #88-29 of M. yessensis, nearly 100 workers of M. yessensis were swarming frantically.Mandibular gripping between workers of the two species was observed, and several M. yessensis corpses were already present, but no corpses of F. lemani.Excavation under that F. lemani nest-opening revealed Table 1 Numbers of surface foraging ants of each species in 1-m 2 quadrats where nest-openings of that species were absent that, in addition to F. lemani workers, there were many large cocoons containing gyne pupae of F. lemani.Consequently, the M. yessensis workers were likely swarming there to plunder immature broods of F. lemani as prey.The nearest nest-opening of M. yessensis was approximately 1.7 m from this location.
The second observation of direct interactions was made in study plot #88-72 on 28 July 1993, 1 month before the overnight study in that plot.At 16:30 h it was first noticed that approximately 25 workers of M. yessensis were flocking around an entrance of an F. lemani nest.Formica workers, which had immature brood in their mandibles, and dealate gynes ran out of the nest-opening and some of them climbed stems of a nearby plant and stayed on the leaves.Manica workers were observed dragging dealate F. lemani gynes and carrying larvae (perhaps of F. lemani) in their mandibles toward their own nest, which was located approximately 0.5 m from the F. lemani nest-opening.This conflict ended by 17:30 h.

DISCUSSION
The present study, using quadrats, was undertaken in late summer.At least during this season, foraging activities of both M. yessensis and F. lemani halted nearly completely during the night.Why both species are inactive at night is unknown, but may be related to low temperature, the difficulty of visual orientation in darkness or large amounts of dew on plants during the night.However, the two species showed an obvious difference in foraging during the daytime.Manica yessensis was crepuscular and F. lemani was continuously diurnal.It may be that Manica avoids strong insolation, because these ants were often observed outside nest-openings under cloud cover, even at mid-day.
The two species showed a distinct difference in activity locations relative to their own nest-openings.Foragers of M. yessensis were strongly confined to the vicinity of their nest-openings compared to those of F. lemani.Under most nest-openings, only a small number of Manica workers were present near the surface.Although the ground between nest-openings of Manica was not dug to investigate their connections, the surface soil was crumbly and the ants were only found immediately beneath the openings.Thus, it is unlikely that nest-openings, which were often separated by as much as 3 m, were connected by underground passages.No immature broods were found in spaces under nest-openings, except for only a few sites.These observations suggest that the nest system of M. yessensis, in which nest-openings were spread over several meters, have functions other than reproduction (brood rearing).The most plausible explanation is that they function as forward bases or outstations for foraging (Traniello & Levings 1986;Anderson & McShea 2001;Cerd a et al. 2002;Debout et al. 2007;Robinson 2014;Sundström 2021).For instance, in Cataglyphis iberica (Emery, 1906), a thermophilic ant inhabiting semi-arid environments of the Iberian Peninsula, Cerd a et al. (2002) demonstrated that polydomy of this monogynous species enhances activities of foragers and retrieval rates of prey (arthropod corpses).Thus, this multiple-nest system could also be effective for M. yessensis to exploit large foraging areas, enabling this ant to retrieve prey items quickly into the nearest nestopenings.Rapid prey retrieval would be important for this rather slow-moving myrmicine ant, because F. lemani, which are much faster than M. yessensis, are abundant in the same microhabitat, and as the present study shows, their foraging periods largely overlap in the daytime.
Similar differences were reported for Myrmica sabuleti Meinert, 1861 and Formica fusca Linnaeus, 1758, which coexist in open grasslands in Europe (Treignes, Belgium), where My. sabuleti was crepuscular, F. fusca was active during the day, and both were inactive at night (de Biseau et al. 1997).In the same area, Deffernez et al. (1990) also reported similar results between two species of Myrmica (schencki Viereck, 1903 and sabuleti) and Formica rufibarbis Fabricius, 1793.At the present study site, high surface temperature (>25 C or even 40 C in full sun) at mid-day, most likely, causes M. yessensis to suspend foraging.At such temperatures, Myrmica species also suspend foraging, even deserting the bait (dead insects) they have dominated, whereas Formica species continue to forage at surface temperatures >40 C (Deffernez et al. 1990;de Biseau et al. 1997).
In a study by de Biseau et al. (1997), distributions of nest-openings of the studied species were not described in detail, but they reported that foraging by My. sabuleti was limited to a radius of 1-2 m from their nests, whereas F. fusca foraged up to 7-8 m from the nest, as observed here between M. yessensis and F. lemani.The present study demonstrated movements of M. yessensis workers between nest-openings more than 3 m apart.When ants forage at new locations, paintmarked workers could be observed at a great distance from where they were marked.Actually, in the survey on 12 September 1993, marked ants were observed foraging in a quadrat 5 m from where they had been marked the previous day.
In the present study site, interspecific relationships, in particular, competition between M. yessensis and F. lemani for various resources (prey, nectar of plants, honeydew of aphids and coccids and nest sites) are presumed to be similar to those reported in studies on Myrmica and Formica (Deffernez et al. 1990;de Biseau et al. 1997).Field experiments using bait, as done in those European studies, were not included in the present study, nor was food retrieval to nest-openings studied.
Another interesting difference between M. yessensis and F. lemani observed in the quadrat study was the frequency with which they utilized plants.Total cumulative numbers from the two surveys recorded 263 F. lemani workers on plants located in the study plot, whereas only one worker of M. yessensis was found on a plant.When ants were found on plants, they were presumably attending aphids.Except for a few instances, aphids were not observed near the ants on plants, and in such instances, we did not explore aphids.In this habitat, however, workers of M. yessensis and F. lemani ordinarily visit floral nectaries and attend aphids on plant leaves, and M. yessensis also attends root coccids underground (K.Masuko, M. Kondoh, unpubl. obs., 1991).Therefore, the reason for the observed difference in their frequencies on plants may be competitive exclusion of M. yessensis by F. lemani, but no further evidence has been obtained.
Finally, if nest-openings are dispersed and food items are carried to the nearest nest, and also if immature broods are raised only in specific nests, then ant colonies must distribute or move food resources between nest-openings.This problem was studied for the meat ant Iridomyrmex sanguineus (Smith, 1858) (Dolichoderinae) by McIver (1991).In this species, multiple nest-openings may be spread over a 400 m 2 area and there are apparent surface trails connecting them.The same problem, although on a much smaller scale, would exist for the nest system of M. yessensis, but this was not investigated in the present study.

Figure 1
Figure 1 Study site at Gotenba Trail New 5th Station (altitude $1,700 m) on Mount Fuji, Japan.(a) Landscape of the study site viewed from the upper point of the ski slope.(b) Quadrat plot (6 m Â 12 m), photographed in May 1994.The vegetation patch marked with an asterisk is centered in quadrat (10, 5) in Figure 3. (c) Three crater-shaped nestopenings of M. yessensis.

Figure 2
Figure 2 Distributions of nest-openings of Manica yessensis and Formica lemani in two nest-areas, (a) #88-74 and (b) #88-47, on Mount Fuji, Japan.Both were investigated on 30 August 1990.Open and solid circles indicate nest-openings of M. yessensis and F. lemani, respectively.Nest-openings of F. lemani were absent in area #88-47.Numerals indicate the number of adult M. yessensis workers found beneath openings by digging.Immature broods (B) and male and gyne alates (S) were also found.Two inset maps show the locations of nest-areas #88-74, #88-47 and surrounding nest-areas.The approximate center of each nest-area is plotted.

Figure 3
Figure 3 Distribution of nest-openings in the study plot (6 m Â 12 m) on Mount Fuji, Japan.All nest-openings of Manica yessensis (circles) and Formica lemani (triangles) in the plot were mapped.Locations of plant stems in four vegetation patches were also mapped (grey dots).Isolated plants that grew alone are not shown.

Figure 4
Figure4Movements of marked Manica yessensis workers among nest-openings of area #88-1 on Mount Fuji, Japan.Ants were paint-marked with different colors at five nestopenings (solid circles) on 1 July 1994; numerals following the color are the number of marked ants.Dashed arrows indicate where and how many marked ants were observed in later censuses.Several marked ants were found in brood chambers (BC) that had no surface openings.A few M. yessensis workers were observed walking between a nest-opening and a thistle (Cirsium purpuratum) (asterisk and Th) and excavating at the base of the plant (observation on 24 July 1994).

Figure 5
Figure 5 Diurnal foraging patterns of Manica yessensis and Formica lemani in the study plot (6 m Â 12 m) on Mount Fuji, Japan.Total numbers of foraging workers of each species in the 72 m 2 area at each census are shown in bar graphs.Surface temperatures were also recorded (lines).(a) Numbers of M. yessensis workers at each census every 3 h from 18:00 h on 28 August to 18:00 h on 29 August 1993.(b) Numbers of F. lemani workers on 28-29 August 1993.(c) Numbers of M. yessensis workers at each census every 2 h from 03:00 h to 19:00 h on 12 September 1993.(d) Numbers of F. lemani workers on 12 September 1993.

Figure 6
Figure 6 Association between observations of foraging ants and nest-openings of Manica yessensis and Formica lemani in 72 quadrats, each 1 m 2 in size, on Mount Fuji, Japan.Quadrats where nestopenings of each species were present are colored yellow.The numeral in each quadrat is the cumulative number of foraging ants during the survey in that quadrat.(a) Frequencies of M. yessensis in the survey on 28-29 August 1993.(b) Frequencies of F. lemani on 28-29 August 1993.(c) Frequencies of F. lemani on 28-29 August 1993, adjusted to the same magnitude as frequencies of M. yessensis.