Colonisation of a Burned Forest by Ants in the Southern Finnish Boreal Forest

Transcription

1 Silva Fennia 30(4) researh artiles Colonisation of a Burned Forest by Ants in the Southern Finnish Boreal Forest Pekka Punttila and Yrjö Haila Punttila, P. & Haila, Y Colonisation of a burned forest by ants in the southern Finnish Boreal Forest. Silva Fennia 30(4): The olonisation of a burned learut by ants in southern Finland was monitored using pitfall traps, artifiial nest sites, and diret nest sampling from the ground and stumps. Clearutting and fire seemed to have destroyed woodant olonies {Formia rufa group), and also other matureforest speies suffered from fire. Myrmia ruginodis was able to survive only in less severely burned moist sites, whereas it benefitted from the enhaned light onditions in a nonburned learut. The fire resulted in an essentially antfree terrain into whih pioneering speies immigrated. The mortality of nestfounding queens appeared to be high. The results supported the hypothesis that the pioneering speies tend to be those that are apable of independent olony founding, followed by speies founding nests through temporary nest parasitism. The suession of the burned learut differed from that of the nonburned one, suggesting that habitat seletion in immigration and priority effets, i.e. ompetition, introdue deterministi omponents in the suessional pathways of boreal ant ommunities. Keywords forest fires, suession, ompetition, ant ommunities, boreal forests Authors' addresses Punttila, Department of Eology and Systematis, Division of Population Biology, P.O. Box 7, FIN0004 University of Helsinki, Finland. Haila, Satakunta Environmental Researh Centre, University of Turku, Reposaari, FIN8900 Pori, Finland Fax pekka.punttila@helsinki.fi Aepted 3 September 996 Introdution thinnings and learutting, whereas the natural disturbanes in primeval taiga are wildfires, storms, flooding, inset outbreaks, and snow dam A major issue in planning onservation of for age (Esseen et ai. 99). Espeially the role of estdwelling organisms is how suessional y fire has been important, and its effets were visiles in managed forests differ from those in more ble in the whole suessional yle of forest natural onditions (Hansen et ai. 99, Hansson stands (e.g. Zakrisson 977, Wikars 99). 99, Haila 994, Haila et ai. 994). In managed Many borealforest organisms have beome forests, the most ommon disturbanes inlude threatened in Finland beause of elimination of 4

2 Silva Fennia 30(4) researh artiles wildfires and promoting ertain management praties like suessive thinnings, planting with onifers, favouring monoultures and shortened rotation times (Heliövaara and Väisänen 98, Rassi and Väisänen 987). Studies omparing natural and managed forest suessions are diffiult to make beause natural suession does not exist or is extremely rare in managed forests. Zakrisson (977) found that about % of forest stands were affeted by fire in his study area eah year in northern Sweden from 55 to 875. In ontrast, nowadays effiient fire suppression has redued the forest area burned to only some hundreds of hetares in Finland eah year (Aarne 995). Further, these oasional fire areas are normally leared and artifiially regenerated. In the seond half of the 950's and the 960's, the mean area prepared for forest regeneration by presribed burning was about 9000 ha in Finland eah year, but sine then also the area of suh burnings has been redued to only a. 700 ha yearly (Aarne 995). Presribed burning might mimi wildfire disturbane with a number of limitations onerning the olonisation and suession of forest organisms. For instane, in presribed burning the learuts are left for drying for some time before the burning and thus, many openountry speies have time to olonise the area (Niemelä et ai. 996). These speies often survive from fire, and they may prevent the olonisation of speialized firedependent speies (Wikars 995). Further, muh smaller amount of surviving, dying and dead trees are left behind in burned learuts as ompared with the situation after wildfires. This may be ruial for many saproxyli invertebrates requiring suh disturbed habitats (e.g. Muona and Rutanen 995, Kaila et ai. 996). However, also presribed burning auses the olonisation, survival and suession of forest organisms to differ from those following mere learutting. If there are suh differenes as regards the disturbane type onerning keystone speies (i.e. speies whih have disproportionately large eologial effets on other organisms; Paine 995), burning may have farreahing onsequenes in the suession and funtioning of the forest eosystem. In boreal forests, keystone speies potentially showing suh differenes in olonisation or survival and suession as regards to disturbane type may be found among trees, myorrhizal fungi, soil mirobes, soil invertebrates, and ants. All these groups inlude speies whih have great impat on a large variety of other organisms and proesses in forest eosystems (for ants, see referenes in Punttila et al. 994). For instane, in boreal ants, the olonisation proess and ommunity suession have been suggested to inlude a number of rather deterministi omponents (Vepsäläinen and Pisarski 98). Thus, priority effets (ompetitive gain from being the first to arrive) in olonisation and early ommunity suession among keystone speies may have longlasting effets in the forest eosystem (Punttila et al. 994, Punttila 996). Studies dealing with the effets of fire on ants and antommunity reovery or suession after fire are numerous from various open fireprone environments, e.g. grasslands, shrublands and heaths (see the reviews by Ahlgren 974, Lyon et al. 978, Warren et al. 987). Although suh studies are also available from many forested environments, e.g. various Australian Eualyptus forests (see the review by Christensen and Abbott 989), different types of tropial forests (e.g. Bentley 976, Morais and Bentley 988, Andersen 99, MaKay et al. 99), and various kinds of oniferous forests (e.g. Heyward and Tissot 936, Pearse 943, Whitford and Gentry 98), knowledge from the Eurasian boreal zone seems to be sare (but see Oinonen 956, Punttila et al. 99). In boreal ant ommunities, Punttila et al. (99) found that among the olonising speies, the firstoming pioneers were those that are apable of independent olony founding. These were followed by speies whih establish new olonies through temporary nest parasitism in the nests of other ant speies. After the area has beome rowded with ant olonies and foraging workers, the only possible mode of olony founding presumably is nest splitting, i.e. budding of the established olonies. In this paper, we fous on three topis: () What are the diret effets of fire on ant ommunities in the boreal forest, () How are the burned areas olonised by ants, and (3) How does the ommunity suession proeed after the fire. 4

3 Punttila and Haila Colonisation of a Burned Forest by Ants... Material and methods We olleted four separate data sets: () A omparison of ant ommunities among learut areas (one area subjet to presribed burning and one nonburned learut of the same age, yrs, in the beginning of the study) and matureforest ontrols in Hyytiälä, northern Häme (for the desription of the study area, see Punttila et al. 99). The areas were monitored for five years (989993) with pitfall traps (plasti ups with a diameter of 65 mm, and a volume of 70 ml, partially filled with 0 % ethylene glyol and detergent, and overed with plasti nontransparent roofs). The trapping was onduted in 4 day periods twie a year (June and August, exept in 989, when both periods were in July) in four sites separated by 5000 m with four sets of 3 traps deposited in a irle with a diameter of a. 4 m. In the matureforest ontrols, only one set of 3 traps was used in eah of the two stands. The monitoring started in the first year after burning. The sampling effort was divided into two periods beause of a seasonal turnover in spider and arabid speies that were also inluded in the study. The ants were identified using the keys of Dlusskij and Pisarski (97), Collingwood (979) and Douwes (98). () The same burned learut was supplied with dark red brik plates (0.5 x.5 x.5 m) to serve as artifiial nest sites mimiking flat stones favourable for ant olonies (Brian 95a, Oinonen 956). The plates were plaed in 6 plots separated by 50 m, eah having 4 sets of 9 plates in a grid with 3 m distanes, the sets being separated by 5 m (6 plates altogether). The plates were deposited in elevated dry loations, and the soil was dug to ensure that the spot was suitable for ants, and shading vegetation was removed (this was repeated during eah inspetion). The plates were heked for their inhabitants late in eah season. In the two first years, an additional heking was done before the nuptial flight of most of the speies in order to reord the winter survival of the queens and olonies, whereas the lateseason hekings were done after the nuptial flight of most of the speies eah year. These data were gathered for four years (98999). (3) We also sampled nests from 4 randomly hosen pine and sprue stumps in two ha squares in the burned learut. New stumps were explored eah year. The stumps were explored arefully and all the soil (some of the stumps were partly overed by soil beause of forestry ativities) and bark on the aboveground part of the stumps were removed, and the revies and soft parts were inspeted. These data were gathered for four years (98999) in the early autumn. (4) In the viinity of the stump samples, we took a random soil sample at m distane from the stump (0.5 x 0.5 m sample square) from whih we searhed ant queens and olonies. The soilsample square was dug up (50 m deep depending on the soil), and all the stiks and ones within the square were explored arefully. These data were gathered for four years (989 99) in the early autumn. We used detrended orrespondene analysis (DCA, Ter Braak 987) to reveal major eologial gradients in the pitfalltrap data. The analysis was performed without transformations for inidene data of ombined athes of the two sampling periods for eah year, and the same weight was given to all speies. Detrending was done by 0 segments. DCA was performed with CANOCO statistial software (Ter Braak 987). 3 Results 3. The ant fauna The pitfalltrap data omprised 647 worker ants and 78 ant queens belonging to 4 and speies (6 speies altogether), respetively (Appendix). The other methods did not yield observations of any additional speies (Table ). The most abundant speies in the study areas aording to the pitfalltrap samples were Formia aquilonia, F. sanguinea, Myrmia ruginodis, Lasius niger, M. sulinodis, and Camponotus heruleanus, whereas the most frequently reorded speies were M. ruginodis (found in 54 of the 300 samples), F. sanguinea (90), L. niger (4), F. aquilonia (4), C. heruleanus (85), and M. sulinodis (7). 43

4 Silva Fennia 30(4) researh artiles Table. Reords of live (lq) and dead (dq) ant queens and olonies () in stumps, artifiial nest sites (plates), and soilsample squares in Stumps Plates Soil M. lobiornis M. ruginodis M. sabrinodis M. sulinodis L. aervorum C. heruleanus L. niger F. fusagroup F. sanguinea lq iq lq lq lq dq iq dq lq Effet of fire on ants In the omparison of burned and nonburned learuts, the burned learut yielded very low athes in the two first study years as ompared with the nonburned learut and the ontrols (Appendix). Of the four speies frequently present in mature forests, i.e. M. ruginodis, C. heruleanus, and the wood ants F. aquilonia and F. lugubris (see Niemelä et ai. 996), only two speies ourred frequently enough to allow statistial testing. The inidenes of both M. ruginodis (X = 47.5, df=,p< 0.00) and C. heruleanus (X = 7.7, df=, 0.00 < p < 0.05) deviated signifiantly from those expeted on the basis of the distribution of the sampling effort among the three treatments. Standardized deviates ((observedexpeted)/expeted) revealed that the inidenes of both speies were unexpetedly low in the burned learut. Thus assuming that prior to the treatments the numbers of these ants were approximately similar in all the study forests, both speies, and espeially C. heruleanus, had suffered from the burning. M. ruginodis seemed to have survived the fire to some extent in moist sites (pers. obs.). Based on the ourrene of olonies of wood ants in lose viinity to the treatment forests (and thus, most likely also in the treatment forests prior to the treatments), the woodant olonies were presumably destroyed by learutting and fire. 3.3 Colonisation of the burned learut The stumps were olonised rapidly by C. heruleanus queens whih started egg laying in small selfmade hambers or tunnels of erambyid larvae under or in the bark (Table ). The numbers of the queens were high in the first years following the fire and rapidly dereased later on during suession. However, the suess of their nestfounding attempts was quite low, as we found many dead queens in the hambers (48 queens, 9 of whih were dead when found, vs 5 surviving olonies). A few suessful queens had established their olonies already in the two first study years (the olonies onsisted of 5 small workers). The size of the olonies grew in subsequent years, and some of them had up to a ouple of dozens workers in the last monitoring year. 44

5 Punttila and Haila Colonisation of a Burned Forest by Ants... The artifiial nest sites experiened a similar olonisation by L. niger queens as did the stumps by C. heruleanus queens (Table ). Most of the olonyfounding attempts were unsuessful under the nest plates (altogether 4 observed L. niger queens one of whih was dead when found, and only five live olonies in the last monitoring year). In the two first years, we heked the artifiial nest sites also in the early season. In 989, no nests were established by July 0, whereas in May 6, 990, two M. sulinodis and one L. aervorum queens were found. None of the L. niger queens present in the late season in 989 were found again. However, we annot exlude the possibility that the queens left the plates and moved into more favourable loations nearby. Unfortunately, we did not use glass plates between the soil and the plate as did Brian (95b) to minimize the disturbane of the monitoring the queens and olonies. Later on, the plates were also olonised by other speies, e.g. four Myrmiines (in addition to F. fusa L.group speies). Espeially M. sulinodis was abundant in the last year monitored. The M. ruginodis olonies, having been quite large, presumably moved under the plates from elsewhere, but in M. sulinodis, all the olonies were inipient ones. In the last year monitored, we found also one inipient olony of the slavemaking ant, F. sanguinea, a speies that employes temporary nest parasitism in olony founding (the olony onsisted of a dozen F. /wsagroup workers and three F. sanguinea workers). The soil samples, on the ontrary, yielded very few queens and olonies (Table ). Only in the third and fourth year we found a notable number of olonies of L. aervorum. This speies has been reported to olonise also hard stumps (Brian 95a). Franh and Espadaler (988) studied also the hard parts of stumps, but only four of the 60 olonies found were loated there. Later on, when the stumps deay, olonies of most speies beome very ommon also in the soil (pers. obs. from nearby older learuts and burned areas). In addition to the nests sampled, we found two F. aquilonia olonies in the burned learut. One rather big mound was loated lose to the edge of the burned learut, and a small olony situated in a moist depression in the burned learut. The latter nest was similar to new bud nests of multinest olonies and we presume that it was a bud nest of the bigger olony (learutting often results in nest splitting in polygynous, i.e. multiqueened, wood ants; Rosengren and Pamilo 978). However, at least the bigger olony had survived beause of ative watering by the employees during the burning proedure (Henrik Lindberg, pers. omm.). The pitfalltrap data revealed that the numbers of queens of M. ruginodis inreased in the burned learut, whereas they were onstantly high in the nonburned learut after the two first years (Appendix). Thus, if ompared with the trends in the numbers of workers, M. ruginodis presumably olonised the burned learut through nest budding, and reahed the density of the nonburned one only in the last study year (Fig., Appendix). Also the unexpetedly high inidene of M. ruginodis workers in the nonburned learut as ompared with the ontrols in the first study year suggest rapid spreading through nest budding following learutting. The ongeneri M. sulinodis, on the ontrary, seemed to have immigrated the burned learut from the surroundings (pers. obs. of many inipient olonies), whereas the nonburned learut remained unolonised by this speies (Appendix, Fig. ). The numbers of queens of C. heruleanus were higher in the nonburned learut than in the burned one. 3.4 Community suession in burned vs nonburned learuts The pitfalltrap athes of many of the speies inreased espeially in the burned learut during the five study years (Fig., Appendix). The DCAordination showed that the ommunities of the open learuts differed from those of the mature forests, and that the ommunity of the burned learut diverged from that of the nonburned one, indiating different suessional pathways between these areas (Fig. ). The burned learut was haraterized by olonisation of M. sulinodis and slightly higher numbers of L. aervorum than in the nonburned learut, whereas F. sanguinea and L. niger prevailed in the nonburned learut already in the first study year (Fig. 3). M. ruginodis was present 45

6 Silva Fennia 30(4) researh artiles 80 M. ruginodis Burned learut G H Nonburned learut Matureforest ontrols 0 80 ill M. sulinodis 80 L. niger H L aervorum F. sanguinea n 80 n C. heruleanus JLJ90 9 F. fusa group H o H o ^ J] Fig. I. The apture frequenies of ommon ants in the burned learut, nonburned learut, and matureforest ontrols in (two sampling periods pooled). Note that the sampling effort in the matureforest ontrols was one fourth of that in the learuts (see Material and methods). in all areas, and it inreased in the burned learut during the study, whereas the wood ants and C. heruleanus were more ommon in mature forests (Fig. 3). 46

7 Punttila and Haila Colonisation of a Burned Forest by Ants. 3 ICM b AE m Bumad learut, a» = Nonbumad learut X Maturaforast ontrols, b A B d C b* A E C B X X D D A E D * I X X X X * B D C EE OC DCA (0.70) Fig.. DCA ordination of the sampling sites aording to their ant samples. Key for symbols is given in the figure. Eigenvalues of the DCA axes are given in parentheses. 3 Lasi nige Form sang Form prat I Myrm rugi Form lema Lept aer*form fuse Myrm sab ~*Myrm sul Myrm lobi Form lugu Form aqui Formrufa Camp here Fig. 3. DCA ordination of ant speies in the sampling sites (parallel to Fig. ). The point for Myrmia lobiornis is indiated by an arrow. 4 Disussion Four main onlusions arise from our results: () Clearutting and fire seemed to have resulted in destrution of the olonies of the oldgrowth dominants, the wood ants, but also other matureforest speies suffered in severely burned areas. () Independent olony founding was very risky for the ant queens beause of the high mortality. (3) There seemed to be a suession of olonyfounding modes in the ants supporting the suggestion of Punttila et al. (99) that the first omers are apable of independent olony founding and that they are followed by speies employing dependent strategies. (4) The ommunities in burned and nonburned learuts diverged from eah other suggesting different suessional pathways. 47

8 Silva Fennia 30(4) researh artiles 4. Effets of fire on ants In many studies dealing with the effets of fire on ants espeially in grasslands and other fireprone environments, it was onluded that the ants often survived the fire, but generally in forested habitats the effets of fire were more severe (Ahlgren 974). In our study, the fire resulted in a largely antfree terrain, with only some M. ruginodis olonies surviving in less severely burned moist sites. Speies nesting in the soil are less affeted than needlemound dwelling wood ants presumably beause the heat does not reah the nest hambers deeper in the soil (Andersen and Yen 985, Neumann 99; see also Vasander and Lindholm 985). Indeed, the fire may gloom for a long time in the remains of woodand mounds (Shimmel and Granström 99, Wikars 99) but sometimes a fration of the inhabitants may be able to survive the burning of the mound (LarsOve Wikars, pers. omm.). In many studies from more fireprone environments an inrease in ant ativity and/or diversity was observed. Suh hanges in ant ommunities have been attributed to () hanges in vegetation and habitat simplifiation (inreased trappability on burned ground as ompared with that in dense vegetation; Majer 980, O'Dowd and Gill 984, Andersen and Yen 985, MCoy and Kaiser 990, Andersen 99, Neumann and Tolhurst 99, York 994), () hanges in resoure levels (redued resoures fore the ants to forage further away from their olonies, or inreased seed fall ativates seedharvesting speies; Majer 980, Andersen and Yen 985, Andersen 988, 99, MCoy and Kaiser 990, Neumann 99, 99, Neumann and Tolhurst 99), and (3) diret and indiret ompetitive effets (ompetitive release when a superior ompetitor suffers disproportionately from the fire; Andersen and Yen 985, Donnelly and Giliomee 985, Andersen 988, 99, MaKay et al. 99). In our data, the lower inidenes of M. ruginodis and C. heruleanus in the burned learut than in the ontrols regardless of the simplifiation of the habitat and ompetitive release (destrution of woodant olonies) presumably impliate true olony mortality aused by fire. Indeed, suh redution in the numbers of M ruginodis was not observed in the nonburned learut. The amount of resoures in the burned learut may have been redued as the treeanopy aphids (an important resoure for wood ants) are absent. On the other hand, some invertebrates may temporarily boost after the fire (Huhta et al. 967). 4. Colony founding The high mortality among nestfounding queens supports the suggestion that the dispersal and nest founding are the most risky enterprises in queen ants' life (Fowler et al. 984, Fowler 987, 99, Tshinkel 987, 99, Deslippe and Savolainen 995 a, 995b, Sommer and Hölldobler 995, Human and Gordon 996). Wilson (97, p. 444) approximated that at most 0. % of virgin queens together with their subsequent offspring an be expeted to survive to olony maturity. In our data, high mortality was observed in the soil nests (in the artifiial nest sites) of L. niger, and in the stump nests of C. heruleanus. However, the longterm persistene of C. heruleanus is questionable in shortliving stumps (Sanders 970). We annot, however, attribute all the disappearanes to queen mortality beause the olonies of both L. niger (Oinonen 956) and M. ruginodis (Brian and Brian 95, Oinonen and Wuorenrinne 976) are apable of hanging loation. The lower survival of queens in the artifiial nest sites than in the stumps may partly have resulted from the disturbane during our inspetion proedure. In our data, the stumps and the artifiial nest sites were the first sites that were olonised by ants, and soil nests beame more ommon later (pers. obs. also from outside the soilsample squares). L. niger and Myrmia speies were the most ommon speies olonising the artifiial nest sites, whereas C. heruleanus and L. niger were the most ommon speies olonising the stumps. Oinonen and Wuorenrinne (976) found that in a Finnish mesi mixed oniferous forest thinned five years earlier 37 % of pine stumps and 7 % of birh stumps were inhabited by ants. The most ommon speies in their data were M. ruginodis (9 of 59 nests enountered in 89 stumps) and L. niger (60 nests), whereas C. heruleanus inhabited only two pine stumps. In a learut mesi sprue forest 54 % of all ant 48

9 Punttila and Haila Colonisation of a Burned Forest by Ants... olonies found were loated in stumps, whereas in a learut subdry pine forest only 9 % of all nests were loated in stumps (Oinonen and Wuorenrinne 976). Although Oinonen and Wuorenrinne (976) attributed the differene to the quiker deay proess in mesi forests, it presumably also owed to the more shading vegetation in mesi forests. In mesi forests the stumps provide more sunexposed onditions than shady ground. Oinonen and Wuorenrinne (976; see also Brian and Brian 95, Oinonen 957) found that the first speies to olonise the hard bark of the stumps was L. aervorum, and to a lesser extent, C. heruleanus. When the bark of the stumps was loose, L. niger and M. ruginodis olonised the stumps, and also F. fusa and F. sanguinea were found inhabiting suh stumps. After the bark had fallen off, there followed an antfree phase whih lasted until the stump was soft. Then they were inhabited again by Myrmia speies, L. niger and F. fusa. The last speies persisting in deaying stumps were M. ruginodis and M. rubra (L.). In southeast USA, Whitford and Gentry (98) found that ants olonised snags only after they were hannelized by termites. In Alberta, Wu and Wong (987) found that 40 % of 49 yrold lodgepolepine stumps were olonised by ants. In their data, it took from 6 to 9 years before ants olonised most of the stumps. In Spain, Franh and Espadaler (988) found that 33 % of oneyear old pine stumps were olonised by ants. The proportion olonised varied between 5000 % in 7 yrold stumps, being highest in 45 yrold stumps. 4.3 Modes of olony founding and suessional pathways In many studies dealing with the effets of fire on ants postfire hanges of the habitat (i.e. suession) were often onsidered to be more important than immediate effets of fire (Ahlgren 974). Our results supported the suggestion by Punttila et al. (99) that in the olonisation proess of antfree terrain there is a suession of immigrating speies with different modes of olony founding. In our data, the firstoming pioneers in the burned learut were those that are apable of independent olony founding, suh as L. niger, C. heruleanus, L. aervorum and M. sulinodis (F. fusagroup speies olonised somewhat later). The olonisation of speies founding their olonies in the nests of other (pioneering) speies through temporary nest parasitism begun later: we found one inipient olony of the slavemaking ant, F. sanguinea, in the last year monitored. In the nonburned learut of the same age the density of M. ruginodis was very high right from the beginning of our study. This supports the idea that M. ruginodis is apable of exploiting the budding strategy. Suh high worker numbers may not have been ahieved through independent olony founding in the short time between the learutting and the beginning of our study (inipient Myrmia olonies grow slowly; Brian 983). An ability to exploit the budding strategy might be benefiial in onditions were the first omers gain ompetitive superiority over later immigrating ones (Punttila et al. 99, 994, Seppä et al. 995). M. ruginodis may be able to monopolize the "Myrrazanihe spae" (sensu Vepsäläinen and Pisarski 98), and thus to prevent the olonisation of eologially similar Myrmiines. The antommunity suession of the burned learut differed from that of the nonburned one. The burned learut was haraterized by the olonisation and growth of olonies of M. sulinodis, steadily inreasing numbers of M. ruginodis, and slightly higher numbers of L. aervorum. The nonburned learut, on the ontrary, was densely inhabited by M. ruginodis already in the first study year, whereas M. sulinodis did not olonise the area during our study. This presumably resulted from habitat differenes between the two learuts, and possibly also from ompetition between M. ruginodis and M. sulinodis. Ant queens seem to be apable of distinguishing suitable nesting habitats, e.g. disturbed areas, in flight for example on the basis of light refletion (Brian 95a, Pontin 960, Brian et al. 966, Wilson and Hunt 966, Fowler 987, Tshinkel 987). The nonburned learut was growing dense Calamagrostis and Deshampsia grasses already in the first study year and thus, the ground was presumably too shady and ool form. sulinodis (Elmes and Wardlaw 98). L. niger and F. sanguinea were present in the 49

10 Silva Fennia 30(4) nonburned learut already in the first study year, and their numbers inreased in subsequent years. These aggressive speies are rather strong ompetitors and they are well adapted to sunexposed onditions. Thus, they are likely to prevent or slow down the olonisation of other aggressive speies (Rosengren et al. 979). Therefore, they are able to rule the ommunity suession in a deterministi manner one they have established themselves in the area (Vepsäläinen and Pisarski 98, Savolainen and Vepsäläinen 988, Punttila et al. 99, 996, Punttila 996). The slavemaking ant, F. sanguinea, seems to be one of the keystone speies in open suessional forest of the taiga. Punttila et al. (996) suggested that ant ommunities in suh forests are organized by diret and indiret effets of ompetition among the top ompetitors, the wood ants and the slavemaking ant. Oinonen (956) found a similar kind of ommunity divergene in open roky forests. He suggested that ompetitive interations resulted in distint ommunities dominated either by F. sanguinea and L. niger or F. fusa. Similar ommunity divergene has been found also in suessional heathlands and sanddune areas in Australia (Fox and Fox 98, Fox et al. 985, Majer 985). Similarly, Andersen (99) found that habitats with different fire regime hosted different ant ommunities, whih was attributable to fireindued habitat differenes and ompetitive interations. Our study area is haraterized by a mosai of young forest stands of various age, habitat type, and treatment history. This variation is refleted in differential olonisation by ants. The ant ommunities in young stands are shaped by various speies interations during suession resulting in a mosai of distint ommunity types (Punttila et al. 996, this study). The type of forestmanagement pratie seems to be a key fator affeting the abundane and distribution of ant speies. Thus this may have farreahing onsequenes in forest eosystems in the long run (Punttila et al. 994). Aknowledgements researh artiles We are grateful to Ilpo K. Hanski and an anonymous referee for omments on earlier drafts of the manusript. We wish to thank Hyytiälä Forest Researh Station for good working failities. The study was finaned by Maj and Thor Nessling Foundation and the Aademy of Finland. Timo Pajunen and Harri Tukia helped in the field work. Referenes Aarne, M. (ed.) 995. Statistial yearbook of forestry 995. Suomen virallinen tilasto. Maa ja metsätalous, 995:5, 354 p. The Finnish Forest Researh Institute, Helsinki. Ahlgren, I. F The effet of fire on soil organisms. In: Kozlowski, T. T. & Ahlgren, C. E. (eds), Fire and eosystems, Aademi Press, New York, p Andersen, A. N Immediate and longerterm effets of fire on seed predation by ants in slerophyllous vegetation in southeastern Australia. Australian Journal of Eology 3: Responses of groundforaging ant ommunities to three experimental fire regimes in a savanna forest of tropial Australia. Biotropia 3: & Yen, A. L Immediate effets of fire on ants in the semiarid mallee region of northwestern Vitoria. Australian Journal of Eology 0: 530. Bentley, B. L Plants bearing extrafloral netaries and the assoiated ant ommunity: interhabitat differenes in the redution of herbivore damage. Eology 57: Brian, M. V. 95a. The struture of a dense natural ant population. Journal of Animal Eology : 4. 95b. Interation between ant olonies at an artifiial nestsite. Entomologist's Monthly Magazine 88: Soial insets. Eology and behavioural biology. Chapman and Hall, London. & Brian A. D. 95. Insolation and ant population in the west of Sotland. Transations of the Royal Entomologial Soiety of London 0:

11 Punttila and Haila Colonisation of a Burned Forest by Ants..., Hibble, J. & Kelly, A. F The dispersion of ant speies in a southern English heath. Journal of Animal Eology 35: 890. Christensen, P. & Abbott,.989. Impat of fire in the eualypt forest eosystem of southern Western Australia: a ritial review. Australian Forestry 5: 03. Collingwood, C. A The Formiidae (Hymenoptera) of Fennosandia and Denmark. Fauna Entomologia Sandinavia 8: 74. Deslippe, R. J. & Savolainen, R. 995a. Colony foundation and polygyny in the ant Formiapodzolia. Behavioral Eology and Soiobiology 37: b. Mehanisms of ompetition in a guild of formiine ants. Oikos 7: Dlusskij, G. & Pisarski, B. 97. Rewizja Polskih gatunkow mrowek (Hymenoptera: Formiidae) z rodzaju Formia L. Fragmenta Faunistia (Warszawa) 6: 454. Donnelly, D. & Giliomee, J. H Community struture of epigaei ants (Hymenoptera: Formiidae) in fynbos vegetation in the Jonkershoek Valley. Journal of the entomologial Soiety of Southern Afria 48: Douwes, P. 98. Hur man känner igen de olika arterna av den vanliga stakmyran. Entomologisk Tidskrift 0: 808. (In Swedish with an English summary.) Elmes, G. W. & Wardlaw, J. C. 98. A population study of the ants Myrmia sabuleti and Myrmia sabrinodis living at two sites in the south of England. II Effet of abovenest vegetation. Journal of Animal Eology 5: Esseen, PA., Ehnström, B., Erison, L. & Sjöberg, K. 99. Boreal forests the foal habitats of Fennosandia. In: Hansson, L. (ed.), Eologial Priniples of Nature Conservation, Elsevier, London, p Fowler, H. G Colonization patterns of the leafutting ant, Atta bisphaeria Forel: evidene for population regulation. Journal of Applied Entomology 04: Patterns of olonization and inipient nest survival in Aromyrmex niger and Aromyrmex balzani (Hymenoptera: Formiidae). Insetes Soiaux 39: , Robinson, S. W. & Diehl, J Effet of mature olony density on olonization and initial olony survivorship in Atta apiguara, a leafutting ant. Biotropia 6: 554. Fox, M. D. & Fox, B. J. 98. Evidene for interspeifi ompetition influening ant speies diversity in a regenerating heathland. In: Bukley, R. C. (ed.), Antplant interations in Australia, Dr. W. Junk Publishers, The Hague, p Fox, B. J., Fox, M. D. & Arher, E Experimental onfirmation of ompetition between two dominant speies of Iridomyrmex (Hymenoptera: Formiidae). Australian Journal of Eology 0: 050. Franh, J. & Espadaler, X Ants as olonizing agents of pine stumps in San Juan de la Pena (Huesa, Spain). Vie Milieu 38: Haila, Y Preserving eologial diversity in boreal forest: eologial bakground, researh, and management. Annales Zoologii Fennii 3: 037., Hanski, I. K., Niemelä, J., Punttila, P., Raivio, S. & Tukia, H Forestry and the boreal fauna: mathing management with natural forest dynamis. Annales Zoologii Fennii 3: 870. Hansen, A. J., Spies, T. A., Swanson, F. J. & Ohmann, J. L. 99. Conserving biodiversity in managed forest. Lessons from natural forests. BioSiene 4: Hansson, L. (ed.) 99. Eologial priniples of nature onservation. Elsevier, London. 438 p. Heliövaara, K. & Väisänen, R. 98. Effets of modern forestry on northwestern European forest invertebrates: a synthesis. Ata Forestalia Fennii 89: 3. Hey ward, F. & Tissot, A. N Some hanges in the soil fauna assoiated with forest fires in the longleaf pine region. Eology 7: Huhta, V., Karppinen, E., Nurminen, M. & Valpas, A Effet of silviultural praties upon arthropod, annelid and nematode populations in oniferous forest soil. Annales Zoologii Fennii 4: Human, K. G. & Gordon, D. M Exploitation and interferene ompetition between the invasive Argentine ant, Linepithema humile, and native ant speies. Oeologia 05: 405^. Kaila, L., Martikainen, P. & Punttila, P Dead trees left in learuts benefit saproxyli Coleoptera adapted to natural disturbanes in boreal forests. Biodiversity and Conservation (In press). Lyon, L. J., Crawford, H. S., Czuhai, E., Fredriksen, R. L., Harlow, R. F., Metz, L. J. & Pearson, H. A Effets of fire on fauna: A stateofknowledge review. U.S. Department of Agriulture, Forest Servie, General Tehnial Report WO6. 43

12 Silva Fennia 30(4) researh artiles MaKay, W.P., Rebeles, A. M., Arredondo, H. C. B., Rodriguez, A. D. R., Gonzalez, D. A. & Vinson, S. B. 99. Impat of the slashing and burning of a tropial rain forest on the native ant fauna (Hymenoptera: Formiidae). Soiobiology 8: Majer, J. D Report on a study of invertebrates in relation to Kojonup Nature Reserve fire management plan. WAIT Department of Biology Bulletin number Reolonization by ants of rehabilitated mineral sand mines on North Stradbroke Island, Queensland, with partiular referene to seed removal. Australian Journal of Eology 0: 3^8. MCoy, E. D. & Kaiser, B. W Changes in foraging ativity of the southern harvester ant Pogonomyrmex badius (Latreille) in response to fire. Amerian Midland Naturalist 3: 3. Morais, de H. C. & Benson, W. W Reolonization of errado vegetation by arboriolous ants after fire. Re vista Brasileira de Biologia 48: (In Portugese with an English summary). Muona, J. & Rutanen, I The shortterm impat of fire on the beetle fauna in boreal oniferous forest. Annales Zoologii Fennii 3: 09. Neumann, F. G. 99. Responses of litter arthropods to major natural or artifiial eologial disturbanes in mountain ash forest. Australian Journal of Eology 6: Responses of foraging ant populations to highintensity wildfire, salvage logging and natural regeneration proesses in Eualyptus regnans regrowth forest of the Vitorian Central Highlands. Australian Forestry 55: 938. & Tolhurst, K. 99. Effets of fuel redution burning on epigeal arthropods and earthworms in dry slerophyll eualypt forest of westentral Vitoria. Australian Journal of Eology 6: Niemelä, J., Haila, Y. & Punttila, P The importane of smallsale heterogeneity in boreal forests: variation in diversity in forestfloor invertebrates aross the suession gradient. Eography 9: O'Dowd, D. J. & Gill, A. M Predator satiation and site alteration following fire: mass reprodution of alpine ash (Eualyptus delegatensis) in southeastern Australia. Eology 65: Oinonen, E. A On the ants of the roks and their ontribution to the afforestation of roks in southern Finland. Ata Entomologia Fennia :. (In Finnish with an English summary) Wood as nest base for ants. Reprint from Metsätietoa /957, Helsinki. 4 pp. (In Finnish with an English summary). & Wuorenrinne, H Uber die finnishen Ameisenarten aus waldpflegerishen Gesihtspunkten. Waldhygiene : Paine, R. T A onversation on refining the onept of keystone speies. Conservation biology 9: Pearse, A. S Effets of burningover and rakingoff litter on ertain soil animals in the Duke forest. Amerian Midland Naturalist 9: Pontin, A. J Field experiments on olony foundation by Lasius niger (L.) and L. flavus (F.). Insetes Soiaux 7: 730. Punttila, P Suession, forest fragmentation, and the distribution of wood ants. Oikos 75: 9 98., Haila, Y., Niemelä, J. & Pajunen, T Ant ommunities in fragments of oldgrowth taiga and managed surroundings. Annales Zoologii Fennii 3: 344., Haila, Y., Pajunen, T. & Tukia, H. 99. Colonisation of learut forests by ants in southern Finland: a quantitative survey. Oikos 6: 506., Haila, Y. & Tukia, H Ant ommunities in learut southern Finnish taiga forests: habitat effets and speies interations. Eography 9: 6 8. Rassi, P. & Väisänen, R. (eds.) 987. Threatened animals and plants in Finland. English summary of the report of the Committee for the Conservation of Threatened Animals and Plants in Finland. Ministry of the Environment, Helsinki. 8 p. Rosengren, R. & Pamilo, P Effet of winter timber felling on behaviour of foraging wood ants (Formia rufagtoup) in early spring. Memorabilia Zoologia 9: 4355., Vepsäläinen, K. & Wuorenrinne, H Distribution, nest densities, and eologial signifiane of wood ants (the Formia ru/agroup) in Finland. O.I.L.B. Bull. SROP, II, 3: 83. Sanders, C. J The distribution of arpenter ant olonies in the spruefir forests of northwestern Ontario. Eology 5: Savolainen, R. & Vepsäläinen, K A ompetition hierarhy among boreal ants: impat on resoure partitioning and ommunity struture. Oikos 5:

13 Punttila and Haila Colonisation of a Burned Forest by Ants... Shimmel, J. & Granström, A. 99. Skogsbränderna oh vegetationen. Skog oh Forskning 4:39^6. (In Swedish). Seppä, P., Sundström, L. & Punttila, P Faultative polygyny and habitat suession in boreal ants. Biologial Journal of the Linnean Soiety 56: Sommer, K. & Hölldobler, B Colony founding by queen assoiation and determinants of redution queen number in the ant Lasius niger. Animal Behaviour 50: Ter Braak, C. J. F CANOCO A FORTRAN program for anonial ommunity ordination by partial detrended anonial orrespondene analysis, prinipal omponent analysis and redundany analysis. TNO Institute of Applied Computer Siene. Wageningen. The Netherlands. 95 p. Tshinkel, W. R The fire ant, Solenopsis invita, as a suesfull "weed". In: Eder, J. & Rembold, H. (eds), Chemistry and biology of soial insets, Verlag J. Peperny, Munih, p Brood raiding and the population dynamis of founding and inipient olonies of the fire ant, Solenopsis invita. Eologial Entomology 7: Vasander, H. & Lindholm, T Fire intensities and surfae temperatures during presribed burning. Silva Fennia 9: 5. Vepsäläinen, K. & Pisarski, B. 98. Assembly of island ant ommunities. Annales Zoologii Fennii 9: Warren, S. D., Sifres, C. J. & Teel, P. D Response of grassland arthropods to burning: a review. Agriulture, Eosystems and Environment 9: Whitford, W. G. & Gentry, J. B. 98. Ant ommunities of southeastern longleaf pine plantations. Environmental Entomology 0: Wikars, L.O. 99. Skogsbränder oh insekter. Entomologisk Tidskrift 3:. (In Swedish with an English summary) Clearutting before burning prevents establishment of the fireadapted Agonum quadripuntatum (Coleoptera: Carabidae). Annales Zoologii Fennii 3: Wilson, E The inset soieties. Belknap Press of Harvard University, Cambridge, Mass. 548 p. & Hunt, G. L. Jr Habitat seletion by the queens of two field dwelling speies of ants. Eology 47: Wu, J. & Wong, H. R Colonization of lodgepole pine stumps by ants (Hymenoptera: Formiidae). The Canadian Entomologist 9: York, A The longterm effets of fire on forest ant ommunities: management impliations for the onservation of biodiversity. Memoirs of the Queensland Museum 36: 339. Zakrisson, O Influene of forest fires on the North Swedish boreal forest. Oikos 9: 3. Total of 79 referenes 433

14 Silva Fennia 30(4) researh artiles Appendix. Total samples (w) and ourrene frequenies (f) of worker ants, and total samples of queen ants (q) in the pitfalltrap data from burned and nonburned learuts (total = periods x 5 traps in eah) and matureforest ontrols (total = periods x 6 traps, 3 traps in two forests) in

15 Punttila and Haila Colonisation of a Burned Forest by Ants... I I I H«THH I I I M ON <N ^H 00 ON I t ^ v o o o o O (N 00 ON ON I I i CO (N O C4 (^ i i rh (N I I I rh rl M» m I oo 00 in I m ON rt I i ( N m m I I I I l o o o o i i i i i M l l i i i g r» I ^ m»h I I I TH I I ^i I l ^ t o ON ^H I N N i I «t ^f H H H m i O N n i9 S J s 3 3 o o H H 435