Nitrogen use strategies of neotropical rainforest trees in threatened Atlantic Forest


Correspondence: Marcos P.M. Aidar, PO Box 4005, São Paulo, CEP 01061–970, Brazil. Fax: 5511 38465763; e-mail:


The characteristics of nitrogen acquisition, transport and assimilation were investigated in species of an Atlantic Forest succession over calcareous soil in south-eastern Brazil. Differences in behaviour were observed within the regeneration guilds. Pioneer species showed high leaf nitrogen contents, a high capacity to respond to increased soil nitrogen availability, a high capacity for leaf nitrate assimilation and were characterized by the transport of nitrate + asparagine. At the other end of the succession, late secondary species had low leaf nitrogen contents, little capacity to respond to increased soil nitrogen availability, low leaf nitrate assimilation and were active in the transport of asparagine + arginine. The characteristics of nitrogen nutrition in some early secondary species showed similarities to those of pioneer species whereas others more closely resembled late secondary species. Average leaf δ15N values increased along the successional gradient. The results indicate that the nitrogen metabolism characteristics of species may be an additional ecophysiological tool in classifying tropical forest tree species into ecological guilds, and may have implications for regeneration programmes in degraded areas.


Although nitrogen is widely recognized as an important ecological factor determining both plant performance and distribution relatively little is known about the characteristics of nitrogen nutrition and metabolism in tropical forest species. In particular we lack information regarding plant N sources in tropical communities since most studies have been carried out in cooler climate ecosystems. It has been suggested that the tropical forest floor contains at least 50% of the nitrogen in the system and that the soil pool size is large relative to annual turnover (Anderson & Spencer 1991). Attiwill & Adams (1993) concluded that up to 25% of the organic matter is mineralized in tropical forests per year and that nitrification is not inhibited in many acidic forest soils. This notion contrasts with previous ideas that nitrification is low in late successional systems due to low soil pH and allelopathic effects on nitrifying organisms (Vitousek et al. 1982). Nutrient dynamics in the Atlantic Forest succession have not been studied in detail. In general, the Brazilian Atlantic Forest soils are highly leached and acid with very low nutrient availability (Leitão Filho et al. 1993). However, we discovered areas of calcareous soil, that differ strongly from other neotropical systems in having higher fertility. Local farmers target these areas for subsistence agriculture (Aidar et al. 2001). Here an area of forest is cut, burned and used for crop growth (‘roça’) for several years without tilling, until productivity declines and a new area is cleared. ‘Roças’ are usually small (<1 ha) and seed dispersal distances are short. During the recovery phase, ecosystem function does not appear to be severely disrupted, and the abandoned ‘roças’ behave similarly to big gaps (Uhl et al. 1990).

Gap-phase dynamics is the fundamental basis for studying the forest growth cycle and it is the beginning of the regenerative cycle. Disturbance created by gaps, that alter environmental conditions, triggers a construction phase characterized by species colonization and growth following a continuum of ecophysiological response that induces cicatrization (Brokaw 1985). Tree species in rain forests can be distinguished according to their successional behaviour into three ecological classes (Budowski 1965; Köhler, Ditzer & Huth 2000). Pioneers and climax species are the extreme positions in a more or less continuous spectrum, and mid-successional species fill the gradient between the extremes. Whitmore (1996) suggests that our current state of knowledge provides evidence that species have broadly overlapping distributions along a gap-size gradient. At our study site (Aidar et al. 2001; Aidar, Godoy & Joly 2003) the early secondary species Piptadenia gonoacantha (Mimosoideae) replaces the pioneer species (e.g. Cecropia spp.). Piptadenia in turn is replaced by late secondary species [Hymenaea courbaril (Caesalpinioideae) and Aspidosperma ramiflorum (Apocynaceae)].

Species from different stages of succession vary in life history and physiological characteristics. Generally, pioneer species have higher photosynthetic capacity and a shorter life span in comparison with species later in succession (Reich, Ellsworth & Uhl 1995). Pioneer species are reputedly high NO3− assimilators, which predominantly reduce NO3 in the shoot, whereas climax species appear to be root N assimilators with an overall preference for NH4+ (Stewart, Hegarty & Specht 1988; Stewart 1991).

In the present study we assessed the patterns of N utilization in a Brazilian Atlantic Forest successional community over a eutrophic calcareous soil (Aidar et al. 2001, 2003) aiming to characterize the strategies involved in tree N acquisition and transport during different phases of forest succession. Such information will contribute to knowledge about neotropical forest biodiversity and regeneration and will help in the development of forest recovery models to support rehabilitation of degraded areas and the management of natural areas.


Study site and plant species

The study area is located in south-eastern Brazil (24°31′43′′ S and 48°41′09′′ W), about 380 km south-west of São Paulo City. Regional climate is tropical hyperhumid without a pronounced dry period. Mean annual precipitation is 1800 mm with 34% of rainfall occurring during the summer and 17% in the winter months. Mean annual temperature is between 17 and 19 °C. Study site description and characterization of the Atlantic Forest succession are detailed in Aidar et al. (2001, 2003).

Selected tree species occurring at three different secondary successional phases after ‘roças’ abandonment were assessed for their primary nitrogen metabolism characteristics. The study included three juxtaposed areas: 15 years old (early succession), 25 years old (mid succession), and a +36-year-old site (late succession). Tree species were chosen according to their relative dominance and occurrence at more than one successional phase. Samples were taken in winter (‘dry’ season; July 1997) and summer (‘wet’ season; December 1997). The classification used (Gandolfi, Leitão Filho & Bezerra 1995) distinguishes three guilds of tree regeneration strategies: (a) pioneer species – specialized in occupation of big gaps, demand light for germination and growth, shade intolerant; (b) early secondary species – specialized to occupy medium-size gaps, showing potential to germinate in shade condition, but demand higher light conditions for growth to reach maturity and/or canopy; (c) late secondary species – specialized to occupy small gaps and understorey, not necessarily demanding light increment for germination and development, can either remain in the understorey or reach the canopy (Table 1).

Table 1.  Characterization of sampled tree species
  1. Key, symbol used to address species; n, number of individuals sampled each season. Occurrence, species occurrence in the different successional phases: early, early succession (15 years old); mid, mid succession (25 years old); late, late succession (+ 36 years old). Leaf pheno., leaf phenology: EG, evergreen; SD, semi-deciduous; DC, deciduous.

Cecropia glaziovi SnethlageCecropiaceaeCgla4PioneerXX EG
Cecropia pachystachya TréculCecropiaceaeCpac2PioneerX XEG
Trema micrantha (L) BlumeUlmaceaeTre3PioneerXX EG
Campomanesia guaviroba (DC) KiaerskowMyrtaceaeCam3Early secondaryXXXEG
Guapira opposita (Vell.) ReitzNyctaginaceaeGua2Early secondary  XEG
Inga marginata Willd.Leguminosae – MimosoideaeIng6Early secondaryXXXEG
Myrcia cf. rostrata DC.MyrtaceaeMyr2Early secondaryXX SD
Piptadenia gonoacantha (Mart.) J.F. Macbr.Leguminosae – MimosoideaePip9Early secondaryXXXSD
Rapanea ferruginea (Ruiz & Pav.) MezMyrsinaceaeRap4Early secondaryXX EG
Schizolobium parahyba (Vell.) S.F. BlakeLeguminosae – CaesalpinioideaeSch2Early secondary  XDC
Tetrorchidium rubrivenium Poepp.EuphorbiaceaeTet3Early secondaryXXXEG
Tibouchina pulchra (Cham.) Cogn.MelastomataceaeTib4Early secondaryXX EG
Aspidosperma ramiflorum Muell. Arg.ApocynaceaeAsp1Late secondary  XDC
Chrysophyllum inornatum Mart.SapotaceaeChr1Late secondary  XEG
Nectandra megapotamica (Sprengel) MezLauraceaeNec1Late secondary  XEG/SD
Ocotea catharinensis Mez.LauraceaeOco2Late secondary  XEG
Roupala brasiliensis KlotzschProteaceaeRou1Late secondary  XDC

Soil N content and availability of soil ammonium and nitrate

The soil characteristics and sampling procedure were described in Aidar et al. (2003), with soil N total content being very consistent throughout the seasons and successional phases with values around 0·15% N. Soil analysis indicated the following averages (and standard deviation; n = 24) for the dry and wet seasons, respectively: pH 5·8 ± 0·4 and 6·0 ± 0·3; organic matter (%): 4·6 ± 1·3 and 5·6 ± 1·5; base saturation (%): 76·3 ± 11·5 and 82·3 ± 9·4; phosphorus (p.p.m): 5·1 ± 2·5 and 3·5 ± 1·6; calcium (p.p.m): 8·4 ± 3·1 and 10·6 ± 4·1. These parameters indicate that the soil can be classified as Cambisols associated with Chernozems (FAO-ISRIC-ISSS 1998) or as Mollisol in the US soil classification (USDA 1998). The turnover of NO3− and NH4+ in the upper horizon, referred to in the following as ‘NO3− and NH4+ availability’, was determined with in situ ion exchange resin bags inserted in the upper 5 cm of the soil following methodology presented in Stewart, Pate & Unkovich (1993).

Leaf assays

Species were sampled from two individuals where possible (65% of individuals), and approximately 10% of individuals were sampled from four replicates. The data were averaged. About 25% of species individuals were sampled from a single individual. The deciduous species A. ramiflorum was sampled only in the wet season.

The in vivo nitrate reductase activity (NRA) measures enzyme activity under non-limiting substrate supply, indicating the amount of active NRA present at the time of sampling and the in vivo substrate-free NRA measures activity based on the available endogenous NO3− pool. Leaf tissue used was collected from the youngest fully developed leaves from branches positioned externally to the crown and they were kept cold until analysis, within 6 h of sampling. Inducible NRA was determined after cut stems were supplied with 10 mm KNO3, under environmental conditions for 24 h. This assay is assumed to measure maximum inducible NRA. All NRA assays were performed as in Stewart et al. (1986).

Fresh leaf blades (0·5 g) from fully developed leaves were cut into pieces and transferred into 5 mL AR grade methanol. After 24 h under ambient temperature, they were frozen until analysis. Leaf NO3− content was determined using 1 mL methanol extracts following Cd-reduction (Sloan & Sublett 1966). Leaves were oven-dried at 60 °C, ground to a fine powder and analysed by continuous flow isotope ratio mass spectrometry (CF-IRMS, Europa Scientific, Crewe, UK) as described by Stewart et al. (1995).

Xylem sap collection and assays

Xylem sap (tracheal sap) was extracted in the morning using a hand vacuum pump (Pate et al. 1994). Two to four shoots with diameters between 5 and 20 mm yielded sufficient fluid, which was kept frozen until analysis. Amino acids and amides in the xylem sap were analysed with a post-column nynhydrin-derivatization high-performance liquid chromatography-based amino acid analyser (Model 6300; Beckman Instruments, Palo Alto, CA, USA). NO3− in the xylem sap was analysed as described above. Ureides (Ur) were analysed using a methodology proposed by Vogels & van der Drift (1970).

Statistical analysis

Data were analysed using the software packages ORIGIN 5·0 (Microcal Software Corp., Northampton, MA, USA) and WINSTAT (R.Fitch Software, Cambridge, MA, USA).


Soil nitrogen content and availability

At all successional stages the availability of soil NH4+ and NO3− was higher in the wet season in comparison with the dry season (Table 2). NH4+ availability was over three-fold higher than NO3− in the dry season in the mid and late successional sites and similar to the availability of NO3− at the early successional site.

Table 2.  Soil N availability determined by resin bag technique
  1. Phase , successional phases: early, early succession (15 years old); mid, mid succession (25 years old); late, late succession (+36 years old). Seasons: dry, July; wet, December. Statistically significant differences (P < 0·05) between groups are indicated by different letters (anova, LSD post-hoc test or t-test): upper case letters (A), within column; lower case letters (a), within row.


Leaf parameters

Pioneer species exhibited the highest NRA (average 660 pkat g−1FW) (Table 3). There was no relationship in these species between NRA and soil nitrate availability. Taking the early secondary species as a whole, the average NRA was markedly lower than for the pioneer species (187 pkat g−1FW) but did show some increase in the wet season (134–239 pkat g−1FW). There is a wide range of NRA among the early secondary species, from less than 20–900 pkat g−1FW. The early secondary leguminous species Inga marginata, P. gonoacantha and Schizolobium parhyba exhibited the highest activities, 448, 236 and 297, respectively. The average NRA for the other early secondary species sampled here was 92. The average NRA of the late secondary species (221 pkat g−1FW) was not significantly different from that of the early secondary species. There was however, a very marked increase in NRA in the wet season (69–342 pkat g−1FW).

Table 3.  Leaf NRA, N content, δ15N and NO3− content for species and successional stages (standard deviation in parenthesis)
  1. Data represent averages of 1–4 leaf samples per species in each season. Successional phases: early, early succession (15 years old); mid, mid succession (25 years old); late, late succession (+36 years old). Species key as in Table 1. ns, not sampled; bdl, below detection limit

PioneerCgladry 727 (7)  2·5 (0·4)  1·13 (0·45)  5 (6) 
wet 504 (85)  2·9 (0·3)  0·43 (0·33)  22 (31) 
Cpacdry 340 (69) 447 (88)1·6 (0·2) 2·3 (0·2)−0·64 (0·93) 1·05 (0·34)4 (6) 1 (2)
wet 289 (10) 919 (140)2·5 (0·3) 2·8 (0·2)−0·45 (0·92) 1·28 (0·42)bdl bdl
Tredry 961 (82)630 2·7 (0·4)2·9  1·44 (0·28) 0·275 (6)6 
wet1019 (0)833 3·3 (0·1)3·1  1·68 (0·11) 0·91bdlbdl 
Early secondaryCamdry 132 (69) 37 2·2 (0·1)1·9 0·23 (0·24)1·6 4 (5)7
wet  65 (39)144 2·5 (0·4)1·9 0·46 (0·01)2·08 17(23)bdl
Guadry  141 (3)  3·9 (0·1)  3·13 (0·47)  84( 58)
wet  355 (473)  4·3 (0·2)  2·38 (0·28)  270 (85)
Ingdry 299 (23)387 (23)215 (69)3·1 (0·1)3·1 (0·0)3·0 (0·5)−0·63 (0·31)1·63 (0·02)2·09 (0·42)29 (41)bdl29 (41)
wet 468 (20)680 (85)641 (252)3·1 (0·0)3·0 (0·2)2·9 (0·2) 0·22 (0·09)1·80 (0·13)1·76 (0·30)bdlbdl7 (9)
Myrdry 125 92 1·41·7  0·641·94  bdlbdl
wet  93 64 1·51·6  1·061·19  bdlbdl
Pipdry  59 (20) 83 (148) 922·8 (0·3)3·2 (0·5)3·1 1·47 (0·54)1·22 (0·59)2·781 (1)2 (3)1·7
wet 900 (23) 54 (37)2273·3 (0·2)3·8 (0·7)4·6 1·37 (0·07)0·69 (0·52)1·64bdl2 (4)56
Rapdry 178 (43)155 (56) 1·9 (0·2)2·4 (0·1)  0·52 (0·69)2·04 (0·32) bdlbdl 
wet  12 (16) 65 (26) 2·4 (0·1)2·2 (0·2)  0·58 (0·37)1·81 (0·11) bdlbdl 
Schdry  218  1·5  2·42  53
wet  375 (157)  2·2 (0·0)  2·29 (0·25)  bdl
Tetdry  51 (39) 56 2·4 (0·1)2·5 1·40 (0·50)2·33 18 (13)5
wet  95 (9) 32 3·3 (0·0)2·6 1·49 (0·95)3·56 4 (6)bdl
Tibdry  40 (16) 46 (65) 1·9 (0·3)2·0 (0·2)  0·91 (0·45)2·09 (0·86) bdlbdl 
wet   0 39 (3) 1·8 (0·1)1·8 (0·2)  1·29 (0·13)1·67 (1·05) bdlbdl 
Late secondaryAspdry  ns  ns  ns  ns
wet  856  2·2  3·11  bdl
Chrdry  175  1·5  4·15  bdl
wet  472  1·6  0·9  bdl
Necdry   74  2·3  1·03  bdl
wet  264  2·4  0·39  bdl
Ocodry    9  3·2  1·68  bdl
wet   42  3·5  1·46  bdl
Roudry   19  1·3  2·63  bdl
wet   74  1·6  3·18  bdl

The average values for the substrate-free assay showed almost no activity, except for the pioneer species Cecropia pachystachya (158 pkat g−1FW) and the early secondary species Guapira opposita (293 pkat g−1FW) in the wet season. The average values for inducible NRA showed a consistent induction for all pioneers in the wet season (Cecropia glaziovi, 1195 pkat g−1FW; C. pachystachya, 1046 pkat g−1FW; Trema micrantha, 1650 pkat g−1FW). Most early and late successional species showed no induction or even a decrease in activity, except for some induction in Tetrorquideum rubrivenium and P. gonoacantha in the wet season (data not shown).

Average leaf N content ranged from 1·27 to 4·87%, with overall average of 2·6 ± 0·7% (Table 3). Average leaf nitrogen contents were highest in the pioneer species (2·66%), a little lower in the early secondary species (2·57%) and lowest in the late secondary species (2·20%). As with NRA there was considerable variation in the leaf nitrogen contents of the early secondary species. Two of the leguminous species had high leaf nitrogen contents (I. marginata, 3·03% and P. gonoacantha, 3·47%). These leaf nitrogen contents were considerably greater than that of the other early secondary species (average value 2·25%). One non-leguminous early successional species, Guapira opposita had a high leaf nitrogen content (4·1%) but the other leguminous species Schizolobium parahyba had leaf nitrogen content of 1·85%. In general leaf nitrogen contents were higher in the wet season (with increases of 21, 11 and 10% for pioneers, early and late secondary species, respectively).

For most species leaf NO3− contents were low or even below the detection limit. However G. opposita had high leaf NO3 concentrations, 84 and 270 µg NO3− g−1FW for the dry and wet season, respectively. In general these low leaf nitrate contents are consistent with the lack of nitrate reductase activity in the substrate-free assay. Guapira opposita did exhibit activity in this assay.

Leaf δ15N values ranged from −1·29 to 4·15‰, with overall average of 1·35 ± 0·97‰ (Table 3). The average value for pioneer species was 0·71‰ compared with 1·53 and 2·06‰ for early and late secondary species, respectively. It is striking that with the exception of one sample, the values of leaf δ15N for the leguminous species were all positive (0·22–2·42‰). Only Inga from the early successional stage had negative values of −0·63 and −0·45% in the dry and wet seasons, respectively. The average δ15N values or the different regeneration strategy guilds and groups indicate a gradual increase as forest succession advances (Table 4).

Table 4.  Average leaf δ15N for regeneration strategy guilds and different groups along successional evolution and the respective linear regressions
  1. Leaf δ15N values were also averaged for species occurring in the different successional phases during both seasons (all dry, all wet). δ15N values for all season were averaged (overall). PS, pioneer species; ES, early secondary species. Successional phases: early, early succession (15 years old); mid, mid succession (25 years old); late, late succession (+36 years old). Statistically significant differences (P < 0·05) between groups within rows are indicated by different letters (anova, LSD post-hoc test); Linear regression, P < 0·05; SE, standard error; n ,number of samples; r, correlation coefficient; a, intercept value; b, slope value.

PS0·51A0·44 80·72A0·1961·16B0·1740·9.87−0·0060·032
All dry0·65A0·25151·34A0·17182·22B0·25140·999−0·4930·075
All wet0·88A0·20151·12AB0·16181·93B0·24140·962 0·0330·050

Xylem sap parameters

The proportion of NO3-N in xylem sap ranged from none (Rapanea ferruginea) to 91% (G. opposita) (Table 5). During the wet season, pioneer species exhibited a 2·5-fold increase in the proportion of NO3 in xylem sap, representing an increase from 0·50 to 2·54 µmol mL−1. Neither the early nor late secondary species showed any seasonal change in either the proportion or amount of nitrate in xylem sap. The concentration of N other than nitrate in the xylem of late successional species was very high in the dry season, representing a 2·6-fold increase when compared with the wet season (21·8 and 8·4 µmol mL−1, respectively). This may indicate reduced transpiration rates of these species in the dry season with a resulting increase in concentration of xylem sap compounds. Xylem sap N composition of the regeneration strategy guilds is shown in Fig. 1. It is evident that for the pioneer species asparagine is the main nitrogen transport compound, comprising nearly 50% of xylem nitrogen. The leguminous early secondary species transport a variety of compounds but in all three species asparagine nitrogen accounted for 24–76% of xylem nitrogen. Djenkolic acid comprised 20–40% of xylem nitrogen in P. gonoacantha. Other early secondary species transported mainly glutamine (45%) and arginine (25%). Guapira opposita transported mainly NO3−, especially during the dry season (89%) and in the wet season asparagine and glutamine were transported. Most of the late secondary species transported asparagine (40–95%) although the proteaceous species Roupala brasiliensis mainly transported arginine (46–67%).

Table 5.  Xylem sap total N content (standard deviation in parenthesis) and main low molecular weight N compounds at different successional stages
  1. Data represent averages of 1–4 xylem sap samples per species in each season. Species key as in Table 1. Successional phases: early, early succession (15 years old); mid, mid succession (25 years old); late, late succession (+36 years old). Xylem sap N compounds listed in order of contribution: Asn, asparagine; Gln, glutamine; NO3, nitrate; Ur, ureides; Djn, djenkolic acid; Ser, serine; Lys, lysine; Arg, arginine; other, other amino compounds.

PioneerCgladry  3·0 (2·0)  Asn (35) NO3 (19) other (19) 
wet 22·2 (24·6)  Asn (45) NO3 (42) other (6) 
Cpacdry 5·9 (4·8)  4·2 (3·7)Asn (50) NO3 (17) other (13) Asn (54) NO3 (7) other (18)
wet12·2 (0·4)  3·3 (0·3)Asn (55) NO3 (33) other (5) Asn (33) NO3 (23) other (22)
Tredry.12·1 (5·4)  3·6Asn (50) unknown (36) NO3 (2)Asn (78) other (13) NO3 (3) 
wet 7·0 (2·9)  6·0Asn (49) unknown (24) NO3 (15)unknown (56) Asn (17) NO3 (6) 
Early secondaryCamdry  7·4 (2·6)  Arg (32) Gln (21) NO3 (18) 
wet  7·6 (1·9)  Gln (31) NO3 (16) Arg (9) 
Guadry   5·5 (0·0)  NO3 (89) Asn (1) Gln (1)
wet   5·1 (0·0)  NO3 (36) Asn (17) Gln (7)
Ingdry 3·3 (2·9) 2·0 (0·7) 2·7 (0·5)Asn (45) Ur (25) NO3 (12)NO3 (33) Ur (18) Asn (16)Asn (40) NO3 (25) Ur (12)
wet 6·2 (0·3) 3·5 (0·2)15·0 (0·8)Asn (63) NO3 (7) Ur (3)NO3 (38) Asn (35) Ur (5)Asn (76) NO3 (3) Ur (3)
Myrdry 9·922·8 Arg (48) Gln (24) NO3 (1)Arg (63) Gln (13) NO3 (5) 
wet  4·3  Arg (43) Gln (23) NO3 (16) 
Pipdry 3·3 (2·9)14·0 (4·2) 9·6Djn (29) Asn (24) Ser (14) NO3 (5)Asn (39) Djn (25) Ser (9) NO3 (2)Djn (38) Ser (14) Arg (10) NO3 (5)
wet 6·2 (0·3)30·7 (20·1)34·7Asn (24) Djn (12) NO3 (15) Ser (11)Asn (28) Djn (21) Arg (12) NO3 (3)Djn (27) Ser (23) Arg (23) NO3 (4)
Rapdry 9·5 (0·7) 3·5 (0·0) Arg (54) Gln (31) NO3 (0)Gln (55) Arg (25) NO3 (0) 
wet 4·5 (2·1) 5·2 (5·9) Arg (36) Gln (47) NO3 (1)Arg (38) Gln (35) NO3 (3) 
Schdry   1·8  Asn (37) NH4 (19) Ur (10) NO3 (9)
wet   1·7 (0·3)  Asn (43) Ur (21) Lys (17) NO3 (2)
Tetdry  8·6 (6·5) 2·5 Gln (55) Arg (18) NO3 (3)Gln (38) Arg (8) NO3 (4)
wet  2·9 1·8 Gln (32) NO3 (7) Arg (4)Gln (28) Arg (12) NO3 (2)
Tibdry 2·8 (1·1) 3·3 (3·1) Gln (65) Arg (10) NO3 (1)Gln (72) NO3 (8) Arg (1) 
wet 1·3 (0·6) 1·5 (0·8) Gln (52) Arg (23) NO3 (1)Gln (68) NO3 (9) Arg (1) 
Late secondaryAspdry  17·2  Asn (62) Arg (19) NO3 (1)
wet   3·7  Asn (39) Arg (23) NO3 (5)
Chrdry  28·6  Asn (95) NO3 (3)
wet   7·3  Asn (91) NO3 (2)
Necdry  31·6  Asn (95) NO3 (2)
wet  15·8  Asn (92) NO3 (2)
Ocodry  20·4  Asn (90) NO3 (3)
wet   8·8  Asn (89) NO3 (0)
Roudry  11·4  Arg (67) Asn (20) NO3 (1)
wet   6·5  Arg (46) Asn (17) NO3 (0)
Figure 1.

Average soluble low molecular weight N compounds composition of the xylem sap (%N) for the regeneration strategy guilds and seasons. NO3, nitrate; Asn, asparagine; Arg, arginine; Gln, glutamine; other, other amino acids; ES, early secondary species.


In contrast to the notion that nitrification is higher in early succession and that ammonification is highest in late succession, we found that NH4+ availability was higher than that of NO3 in the early successional phase during both the wet and dry seasons. No differences in NH4+ and NO3− availability were found at the late successional stage whereas wet season nitrate availability was 1·8 times that of ammonium at the mid-succession stage. The lower rate of nitrification seen in the early stage may relate to nutrient exhaustion of the soil following cultivation.

However, soil N levels did not change along the studied succession after farming for 3–4 years without fertilizer application. Our results contradict studies of soil inorganic N in the Amazon region (Neill et al. 1995; Verchot et al. 1999) who found that soil N availability is lower in secondary forests compared with mature forests, and this is likely to be attributable to the nutrient-rich calcareous soil at the study site.

The higher nitrate reductase activities observed in pioneer species compared to activities of early/late succession species are consistent with other published work (Havill, Lee & Stewart 1974; Lee & Stewart 1978; Stewart & Orebamjo 1983; Smirnoff, Todd & Stewart 1984; Stewart et al. 1988, 1990). The greater fertility of this site compared with others that we have examined in Brazil (Stewart, Joly & Smirnoff 1992) is seen in the considerably higher nitrate reductase activities exhibited by early and late secondary species.

The relatively high NRA of the leguminous early secondary species suggests that they are active in nitrate assimilation and in both wet and dry seasons. Consistent with this is the observation that xylem nitrate concentrations were higher than those of other early secondary species (G. opposita excepted). Two of these species, Inga and Piptadenia had amongst the highest leaf nitrogen contents of species studied here. The other leguminous species Schizolobium had appreciably lower leaf nitrogen 1·85% N compared with 3·03 and 3·47% for Inga and Piptadenia. The average leaf δ15N value for Schizolobium was 2·36 compared with 1·15 and 1·53 for Inga and Piptadenia. The higher leaf nitrogen content and the lower leaf δ15N values of Inga and Piptadenia suggests the possibility of some N input from nitrogen fixation in these two species. Many species belonging to Leguminosae typically have nitrogen-rich leaves, which imposes a high N requirement and consequently when N is available, tend to maximize N uptake, increasing overall N content. The advantages of this lifestyle may only be realized in particular habitats with relatively rich soils, where ambient temperature, light and water availability are high (McKey 1994), as occurs at the study site. Australian members of the Mimosoideae appear to have highly diverse nitrogen-use strategies, with an ability to fix N2, access organic N via mycorrhizal associations and to assimilate NO3− and NH4+ (Schmidt et al. 1998; Stewart & Schmidt 1999). These two Brazilian species may assimilate both nitrate and N2.

The product of N2 fixation is transported in the xylem sap as amides or as ureides, the latter being restricted to certain taxonomic groups (Giller 2001). However, ureides are produced by many non-fixing plants and their presence is not sufficient to prove that N2 fixation is occurring (Peoples et al. 1991; Herridge et al. 1996). In accordance with this suggestion the presence of ureides in the xylem sap of S. parahyba is likely not to be indicative of N2 fixation since this species has been reported to be non-nodulated (Barberi et al. 1998). Schizolobium parahyba had the highest ureide concentration during the wet season when it is producing flowers and fruit. Ureides detected in the xylem sap of S. parahyba in the wet season could originate from N storage compounds formed during degradation of nucleic acids in senescent tissues which are subsequently transported to growing tissues (Peoples et al. 1991). In contrast, the presence of ureides in the xylem sap of P. gonoacantha and I. marginata is a possible indicator for N2 fixation since both are considered to be nodulating species (de Faria et al. 1989; de Faria & de Lima 1998) and active nodules were detected in the field (Aidar, personal obs.).

Many of the non-leguminous early and late secondary species that have low leaf N contents belong to families such as Myrtaceae, Melastomataceae and Proteaceae (Vitousek & Sanford 1986; Foulds 1993; Ellsworth & Reich 1996; Nogueira & Haridasan 1997; Schmidt & Stewart 1998; Stewart & Schmidt 1999). It is of interest that species belonging to these families are very common in early and late successional stages in Atlantic Forest over oligotrophic acidic soil, with Tibouchina pulchra (Melastomataceae) being the dominant tree species in the early secondary succession. These observations suggest the possibility that the domination of Leguminosae species in the secondary succession at our study site is a consequence of the richer soil (eutrophic calcareous soil), in which the higher nutrient availability, especially N, favours the success of these species over those adapted to oligotrophic substratum (McKey 1994; Aidar et al. 2001, 2003).

The average δ15N value of pioneer species was 0·71‰ whereas those of early and late secondary species were 1·53 and 2·09‰, respectively. These results contrast with those obtained by Pate, Stewart & Unkovich (1993) who found that high δ15N values were related to utilization of NO3. Increasing δ15N values of species across the succession contrast with declining nitrate reductase activities, with values for pioneers of 660 pkat g−1FW, with those for early and late succession species being 187 and 221 pkat g−1FW, respectively. Increasing δ15N values along a successional gradient were also observed in Amazonian Forest (Thielen-Klinge et al. 1999) suggesting changes in the N cycle throughout the succession. The 15N natural abundance integrates N cycle processes (Robinson 2001), and it would be interesting to compare the δ15N values of species growing at both calcareous and acidic Atlantic Forests.

The high concentration of asparagine in the xylem of pioneer species suggests that although these species appear active in leaf nitrate assimilation there must be substantial root nitrogen assimilation. The only early secondary species to show appreciable nitrate in the xylem was G. opposita, with concentrations comparable to those of pioneer species. The relatively low concentrations of other nitrogenous compounds in the xylem of this species suggest that it may have little root nitrogen assimilation.

The profile of nitrogenous compounds in xylem sap was characteristic for the different regeneration strategy guilds, and was relatively constant between seasons. Across the succession there appears to be a shift from pioneer species transporting nitrate + asparagine, to early secondary species transporting glutamine + arginine and the late secondary species transporting asparagine + arginine.

The data presented here indicate that the nitrogen nutrition strategies of pioneer and late secondary species are readily distinguishable. The pioneer species have a high capacity for nitrate assimilation, are able to respond to increased availability of soil nitrogen, have high leaf nitrogen contents and transport nitrogen predominantly in the form of asparagine and nitrate. The late secondary species have a lower capacity for nitrate assimilation, are less able to respond to changes in nitrogen availability, have lower leaf nitrogen contents and transport nitrogen in the form of asparagine and arginine. The nitrogen characteristics of pioneer species appear to match their ecological characteristics fairly well in that they are adapted to rapid growth in open areas (forest gaps), where they can maximize carbon gain through a high photosynthetic capacity (Chazdon et al. 1996), which in turn, is highly dependent on nutritional status (Reich & Walters 1994). The late successional species have slower growth rates, a longer life cycle and lower assimilatory capacity. The occurrence of arginine as a major nitrogen transport compound may reflect greater internal storage and remobilization of nitrogen (Millard 1995) than is seen in pioneer species.

The early secondary species studied here form a less homogeneous group than the pioneer or late secondary species. The two Mimosoideae show some of the characteristics of the pioneer species. They have high leaf nitrogen contents, a capacity for leaf nitrate assimilation and can respond to increased soil nitrogen availability. At the other end of the spectrum are species such as Myrcia cf rostrata and Tibouchina pulchra that resemble the late secondary species in having low leaf nitrogen contents, low capacity for leaf nitrate assimilation, little capacity to respond to increased soil nitrogen availability and employ arginine as a nitrogen transport compound. Our observations on the nitrogen use strategies of different species would seem to suggest that there is a continuum of species across the forest succession with pioneers and late secondary species representing the extremes of this continuum.

These results indicate that the strategies involved in the acquisition and transport of N throughout forest succession and seasons may be an additional tool for the diagnosis of the tree's regeneration strategy guilds (ecological guilds or classes) in the neotropical forest (Fig. 2).

Figure 2.

Conceptual diagram of the forest succession showing the main low molecular N-compounds transported by the different regeneration strategy guilds and the observed trends of analysed parameters. Leaf δ15N (‰); xylem sap N content (µmol mL−1); NRA (pkat g−1FW); leaf N content (%); xylem sap NO3 (%); NO3, nitrate; Asn, asparagine; Arg, arginine; Gln, glutamine; other, other amino acids. Early succession – 15-year-old-after crop field abandonment; mid succession – 25 years of abandonment; and late –+ 36 years after abandonment.

The characterization of the Atlantic Forest over calcareous soil found at the study site is a important contribution to the knowledge of Neotropical Biodiversity once considering that it is probably the only site where it occurs in Brazil, being a very important genetic bank of threatened species such as Ocotea catharinensis Mez – Lauraceae; Aspidosperma ramiflorum M.Arg Apocynaceae; Myrocarpus frondosus Allemao – Fabaceae; Chrysophyllum inornatum Martius Sapotaceae; and Euterpe edulis Mart. Arecaceae (;Joly et al. 1999; Aidar et al. 2001). Only 8% of the original coverage of Atlantic Forest remains today (SOS MATA ATLANTICA 1998) making Atlantic Forest one of the most threatened tropical forests, as it is also considered to be one of world's biodiversity ‘hot spot’ (Myers 1988; Myers et al. 2000). The characterization of this locally restricted successional pattern is decisive in supporting actions taken to rehabilitate degraded lands, and can be a potential tool for sustainable tropical forest management, including the potential to serve as a carbon sink (Silver, Ostertag & Lugo 2000; Aidar et al. 2002) in this highly endangered bioma


This project was founded by CAPES/MEC, which granted an international fellowship to M.P.M.A. M.P.M.A. thanks Marcos S. Buckeridge for the continuos encouraging support; Ian Biggs, Ms. Lisha Allen, Nicole Robinson and Tanuwong Sangtiean for their friendship and support in the Plant Metabolism Laboratory, University of Queensland; MSc. João R.L. de Godoy, Orlei Lopes, José and Eufrásio da Mota for their support in the field work.

Received 23 April 2002; received in revised form 22 July 2002; accepted for publication 7 August 2002