Litter as seedbed: interactions between the soil, seedlings and litter of kauri ( australis)

Strooisel als zaaibed: over de interacties tussen de bodem, zaailingen en strooisel van kauri ()

Promotor Prof. dr. F. Berendse Hoogleraar in het Natuurbeheer en de Plantenecologie, Wageningen Universiteit

Copromotor Dr. W.G. Braakhekke Universitair docent, leerstoelgroep Natuurbeheer en Plantenecologie, Wageningen Universiteit

Samenstelling promotie- Prof. dr. A.M. Cleef commissie Universiteit van Amsterdam

Prof. dr. ir. G.M.J. Mohren Wageningen Universiteit

Prof. dr. ir. N. van Breemen Wageningen Universiteit

Prof. dr. M.J.A. Werger Universiteit Utrecht

Dit onderzoek is uitgevoerd binnen de Nederlandse onderzoeksschool SENSE (Research School for the Socio-Economic and Natural Sciences of the Environment). Litter as seedbed: interactions between the soil, seedlings and litter of kauri (Agathis australis)

Eric Verkaik

Proefschrift ter verkrijging van de graad van doctor op gezag van de rector magnificus van Wageningen Universiteit, Prof. dr. M.J. Kropff, in het openbaar te verdedigen op woensdag 6 december 2006 des namiddags te half twee in de Aula.

Verkaik, E. 2006. Litter as seedbed: interactions between the soil, seedlings and litter of kauri (Agathis australis). PhD-thesis Wageningen University, with a summary in Dutch.

ISBN 90-8504-549-5

The research for this thesis was carried out at the Nature Conservation and Ecology Group, Department of Environmental Sciences, Wageningen University, The Netherlands.

Cover photo: view from the bridge crossing the Waitakere stream, at the Auckland City Walk. In the centre of the photograph a kauri tree. Abstract

Verkaik, E. 2006. Litter as seedbed: interactions between the soil, seedlings and litter of kauri (Agathis australis). PhD-thesis, Wageningen University, Wageningen, The Netherlands. ISBN 90-8504-549-5. 112 pp.

Plants have important impacts upon soil processes such as nutrient mineralisation and dynamics. They might even enhance their own fitness by improving soil quality or by making the soil less favourable for competing species. In the latter strategy, in plant foliage might be important. To investigate this strategy we used the kauri tree (Agathis australis), whose massive litter accumulation has a huge impact upon the soil, as an example. We hypothesised that kauri-tannins can reduce mineralisation in the soil below the crown of a kauri tree and thus create an environment where kauri seedlings are better able to compete with angiosperm seedlings, resulting in natural selection for kauri trees with high concentrations. To test this hypothesis, several descriptive and experimental studies were conducted. Nutrient concentrations in the foliage of a common shrub species show that below the crown of a kauri tree nutrient availability of the soil is lower than outside the kauri crown. Two studies on tree seedlings in kauri forest indicate that the lower nutrient availability below kauri reduces the competition for kauri seedlings from fast growing angiosperm seedlings. The results of a laboratory incubation study with purified kauri tannins suggest that kauri tannins contribute to the reduction of nitrogen availability below kauri by complexing proteins. A litterbag experiment provided no conclusive evidence in favour or against this effect of kauri tannins on nitrogen availability. Finally, a study on the variation in tannin concentration between kauri seedlings shows that the variation in tannin concentration of kauri foliage is partly genetic. Therefore natural selection of kauri individuals with high tannin concentration is possible. Overall the results of the different studies confirm the main hypothesis: kauri tannins can reduce nitrogen mineralisation in the soil under the crown of a kauri tree and thus create an environment where kauri seedlings are better able to compete with angiosperm seedlings. Natural selection for kauri trees with high tannin concentration, via the effects these tannins have upon nitrogen availability and thereby on the competition between seedlings, is plausible.

Key words: competition, ecosystem engineer, kauri (Agathis australis), nitrogen mineralisation, plant–soil interactions, tannins Dankwoord - Acknowledgements

Uiteindelijk komt er dan toch een einde aan dit promotieproject. Er waren momenten dat ik de handdoek in de ring wilde gooien. Zoals bijvoorbeeld in 2004 toen het opzuiveren van de tanninen maar tijd bleef kosten waardoor ik mijn laatste veldwerk een half jaar moest verschuiven. Maar er waren natuurlijk ook veel mooie momenten. Zo denk ik bijvoorbeeld met veel plezier terug aan het kauribos met al die vreemde beesten en planten of aan het schitterende uitzicht in ‘little ’. Ik heb dit promotieproject zeker niet alleen gedaan en er zijn veel mensen te bedanken die dit werk mogelijk hebben gemaakt. Een aantal zal ik hier bij naam noemen. Ten eerste mijn begeleiders Frank Berendse en Wim Braakhekke. Frank, jij vertrouwde me deze klus toe en gaf me veel kritisch commentaar bij het opzetten van de experimenten. Vooral je roep om herhalingen ben ik in de loop van de tijd erg gaan waarderen. Verder corrigeerde je op een kritische manier al mijn schrijverij. Wim, je werd pas later bij het project betrokken en ik waardeer het erg dat je op dat moment veel tijd nam om het onderzoeksplan van A tot Z te doorlopen. Je liet me daardoor ook zien dat de oorspronkelijk vraagstelling van het onderzoek nog steeds relevant was. Daarnaast droegen je vele kritische opmerkingen zeker bij aan de kwaliteit van de verschillende hoofdstukken. Het oorspronkelijke idee voor een onderzoeksproject naar de invloed van kauri op bodem en vegetatie komt van Nico van Breemen. In het begin van het project kwam onze kauri-projectgroep regelmatig bij elkaar. Vanuit de leerstoelgroep Bodemkunde en Geologie waren naast Nico van Breemen vooral Peter Buurman, Toine Jongmans en Tom Veldkamp bij dit project betrokken. De discussies in de vergaderzaaltjes hebben me meer geleerd over bodem en geologie en daarnaast was het leerzaam maar ook gezellig om samen door Nieuw- Zeeland te reizen. Kauri-AIO’s Lieven Claessens en Anne Jongkind, ik vond het leuk om met jullie samen te werken. Vooral het primitieve wonen in Titirangi en Henderson was gezellig, maar het was ook goed dat we ervaringen konden delen als het onderzoek soms tegenzat. De afstudeervakstudenten Evelien Vonk en Marijn Zwart verrichtten veldwerk in Nieuw-Zeeland en verzamelden zo op enthousiaste wijze data voor dit onderzoek. In Wageningen werkte ik bij de leerstoelgroep Natuurbeheer en Plantenecologie, waar de sfeer in de barak, op enkele verhitte discussies na, gemoedelijk was. Veel collega-AIO’s en ook postdocs waren daar in de loop van de tijd voor verantwoordelijk. Spelletjes-, film- en fondue-avonden, maar ook vogelexcursies droegen bij aan de goede sfeer. Ook ben ik een aantal van jullie dank verschuldigd voor tips over metingen en statistische analyses. Mijn kamertje in de barak deelde ik met Gabriela Schaepman, die met stimulerende maar ook relativerende opmerkingen zorgde voor een prettige werksfeer. Ook de vaste staf en het ondersteunend personeel droegen bij aan de prettige en productieve sfeer. Frans Möller wil ik speciaal noemen, voor het vele maalwerk dat hij heeft verricht en voor de goede verzorging van de kaurizaailingen in de kas. Bij Jan van Walsem wist ik mijn vele plant- en bodemmonsters in goede handen. Hij hielp daarnaast met goede raad bij de lastige klus van het opzuiveren van tanninen. Ook wil ik het secretariaat bedanken voor hun hulp en dan met name Marriette Coops en Gerda Martin. Gerdien de Jong van de Universiteit Utrecht gaf advies over kwantitatieve genetica, en Joyce Burrough gaf advies over het Engels in de hoofdstukken vijf en zes. Besides the help in the Netherlands I also got a lot of support in New Zealand. First of all I would like to acknowledge botanist Rhys Gardner! Rhys, your help and hospitality really made a difference. I had very good times staying at your place. Not only because of the good food, but also because of the good company and conversations. At times, your house was like a field station, with my plant material not only in the garage but also in the living room. You learned me the names of many plant species, helped me out with small and almost undeterminable seedlings and provided me with much information on the Waitakere Ranges. You also showed me some nice places in and around Auckland. Furthermore, I appreciated your help with establishing the seedling experiment. The Evans family of the Tree House at Kohukohu provided help during our first visit to New Zealand, Suzanne Koster took me on a nice trip to Tongariro, Warwick Silvester and Bruce Burns provided support and information on kauri forests, and Tim Lovegrove of the Auckland Regional Council helped us to obtain work-permits and housing. The Rangers of the Auckland Regional Council and the caretakers of the dams of Watercare provided help too, while the Auckland Regional Council, the New Zealand Department of Conservation and Mr Lee gave permission to work in their forest areas. In addition, the Auckland Regional Council provided rainfall data. Verder wil ik mijn ouders noemen, van wie ik zoveel geleerd heb en die ook deze periode een steun voor mij waren. De rest van mijn familie en de familie van Janneke waren mij ook tot steun gedurende deze periode, door het bieden van een luisterend oor. Janneke, jij tenslotte was een grote steun voor mij. Het was fijn om met iemand over allerlei aspecten van dit project te kunnen praten. Verder waardeer ik het begrip dat je opbracht, vooral ook tijdens de laatste loodjes van het project.

Allen bedankt! Thanks to all of you!

Eric, Wageningen, september 2006.

Table of contents

Chapter 1 General introduction 10

Chapter 2 The effect of kauri trees on the availability of nutrients, water and light 20

Chapter 3 Seedling distribution below and outside the crown of kauri trees as 34 affected by site conditions

Chapter 4 The decrease of the soil resource by New Zealand kauri trees increases 46 the relative fitness of their seedlings

Chapter 5 Short-term and long-term effects of tannins on nitrogen mineralisation 58 and litter in kauri forests

Chapter 6 Variation in tannin concentration among species of New Zealand kauri 70 forest and the effects of kauri tannins on litter decomposition and nitrogen mineralisation.

Chapter 7 General discussion 84

References 94

Samenvatting 102

Summary 106

Curriculum vitae 110

Chapter 1

10 General introduction

Chapter 1

General Introduction

‘The mature tree is often more than 35 m tall. It is only in the forks of large branches that any epiphyte can occasionally retain a hold. Besides shedding copious amounts of bark, the large trees are ‘self-pruning’; that is, any small branches that are heavily shaded are cut off by the plant, and fall to the ground. The shedding of older is also a regular process. All this debris piles up beneath the kauri tree and is quite slow to rot, so that the forest floor around it is covered by a thick layer of litter – called pukahukahi or pukahu by the Maori, and bookau by the bushmen. The in this litter of dead tissue seems to retard the growth of various organisms, and many seedlings find difficulty in becoming established; as a result, the interior above the shoulder-high kauri grass is more open than are other parts of the forest. This is one of the things that makes the kauri grove so impressive, the visitor pushes his way through dense undergrowth which suddenly opens out to reveal the great trunks before him.’ (Bieleski 1979).

11 Chapter 1

A positive feedback between trees and the soil

Trees have important impacts upon soil processes such as nutrient mineralisation and organic matter dynamics (e.g. Binkley 1995). The effects trees have upon the soil differ between species, and as a result different tree species growing on the same site can lead to different soil characteristics (Binkley 1995; Prescott 2002; Wardle 2002). A study in a mixed forest in the north-eastern United States for instance, shows how calcium availability in the soil differs below two tree species as a result of differences in nutrient uptake and litter quality (Dijkstra 2001). Similar results were found in mixed beech-spruce forest in Austria (Berger et al. 2004). The different effects species have upon the soil can also lead to differences in the understory vegetation (Hommel et al. 2002) and might affect the fitness of the individual trees themselves. Frelich et al. (1993) studied a forest mosaic consisting of patches dominated by hemlock (Tsuga canadensis) or sugar maple (Acer saccharum), and found these patches to result from the influence these species have, via the soil, upon each other’s seedlings. A positive feedback can develop between trees and soil, whereby the trees affect soil properties and the changes in the soil properties have positive effects on the fitness of the trees (e.g. Hobbie 1992; Berendse 1994a; Van Breemen and Finzi 1998; Clark et al. 2005). Reviewing the literature Binkley and Giardina (1998) found examples of tree species that have a positive effect upon their own fitness as well as examples of tree species that have a negative effect upon their own fitness due to their effects on soil processes.

Polyphenols and tannins

Polyphenols and especially tannins might play an important role in a positive feedback between plant and soil. Plant species adapted to nutrient-poor soils have a number of characteristics in common, such as a relatively large root , a low growth rate and a low nutrient loss rate. Leaves of growing on nutrient-poor soils commonly have high concentrations of polyphenolic substances (Chapin 1980; Kuiters 1990; Hobbie 1992). Phenolics are chemically defined as compounds that possess an aromatic ring bearing a hydroxyl substituent, including their functional derivatives (Figure 1.1). Polyphenols have several or many phenolic hydroxyl substituents. The most abundant polyphenols of woody plants are condensed tannins. Tannins are polyphenolic compounds ranging in molecular weight between 500–3000 Daltons (Hättenschwiler and Vitousek 2000; Kraus et al. 2003a). Polyphenols, and especially large molecules like tannins, can complex proteins through hydrogen bonding and hydrophobic effects (Figure 1.1). Thereby they can prevent the

12 General introduction

breakdown of plant proteins by fungi and bacteria and also inhibit essential digestive enzymes of herbivores. It is assumed that high polyphenol concentrations reduce biomass loss caused by pathogens and herbivores, and thereby also reduce the annual nutrient requirement of plants (Chapin 1980; Schultz et al. 1992; Van Genderen et al. 1996; Harbone 1997). High polyphenol concentrations of can reduce the rate of litter decomposition, resulting in the accumulation of a thick humus and litter layer (mor humus). It is further suggested that high polyphenol concentrations results in a low rate of nitrogen mineralisation in the soil (Kuiters 1990; Field and Lettinga 1992; Hättenschwiler and Vitousek 2000; Kraus et al. 2003a).

ROHOC R H bond

phenol protein

Figure 1.1. Hydrogen bonding of a phenolic compound with a protein. The hydroxyl group of the phenolic compound, which is attached to the aromatic ring, forms a bond with the carbonyl group of the protein. After Field and Lettinga (1992).

There are various mechanisms by which high concentrations of polyphenols in plant litter can feedback to the plants positively via the litter layer and the soil. Firstly, a reduction of nitrogen mineralisation in the soil, due to high polyphenols levels in plant litter, might make the soil less favourable for competing species which need higher levels of available nitrogen. Studies in heathland vegetation, for instance, have shown that plant species affect the plant succession via the litter they produce. Ericaceous species, which are dominant early in the succession, produce litter that decomposes slowly thus keeping conditions poor for a long time. The perennial grass species that follows the Ericaceous species in the succession, can only establish when the soil has become sufficiently fertile. This grass species produces litter that decomposes more easily, thus increasing the nitrogen mineralisation rate and creating better conditions for the growth of grass species (Berendse 1994b; Berendse 1998). On river floodplains in Alaska high concentrations of tannins and phenolic compounds might help

13 Chapter 1

poplar to gain dominance over alder, by reducing nitrogen mineralisation in the soil (Schimel et al. 1996). Other ways in which polyphenols in plant litter can benefit the plant via the soil are by changing the availabilities of dissolved organic nitrogen (DON) and inorganic nitrogen (Northup et al. 1995; Northup et al. 1998), by reducing nutrient losses, or by reducing the toxic effects of aluminium in acid soils (Hättenschwiler and Vitousek 2000; Kraus et al. 2003a). The effects of plant induced soil changes on the plant’s fitness have not been extensively studied, and the role of tannins in such a feed back is still unclear (Van Breemen and Finzi 1998; Hättenschwiler and Vitousek 2000; Kraus et al. 2003a).

Research program ‘Podzolisation under kauri, for better or worse?’

For a number of reasons the New Zealand kauri tree (Agathis australis (D. Don) Lindl.) is an interesting species for studying a positive feedback between a tree species and the soil mediated by tannins. Kauri belongs to the order of the Coniferales and the family of the . This family comprises 40 species from three genera: Araucaria, Agathis and Wollemia, and is one of the oldest existing families (Gifford and Foster 1989; Hill and Scriven 1998; Brophy et al. 2000; Lovis 2001; Eagle 2002). The majority of these species are large, often emergent, trees in rain forests of the southern hemisphere (Enright et al. 1999). Kauri belongs to the genus Agathis which is found in montane tropical forests in Australia, Papua New Guinea, New Caledonia and elsewhere in the Pacific, and which reaches its southern limit in New Zealand (Beveridge 1975; Whitmore and Page 1980; Ogden and Stewart 1995). Kauri occurs naturally at the of New Zealand, and is a large tree, up to 30 m or rarely up to 60 m tall, with a columnar trunk up to 3 m, exceptionally up to 7 m, in diameter (Sale 1978; Ecroyd 1982). The normal attainable age of kauri is 600–700 years (Ahmed and Ogden 1987). Under mature kauri trees very thick organic layers may develop, reaching 2 meters or 2 more in depth near the stem, with a biomass of up to 55 kg/m accumulated in this layer (Yeates et al. 1981; Silvester and Orchard 1999). This organic layer contains large amounts of 2 nitrogen, up to 0.65 kg N/m , that seem inaccessible to plants (Silvester 1978; Silvester 2000). It appears that the thick organic layer leads to podzolisation of the underlying soil (Swindale 1957; Beveridge 1977; Wells and Northey 1985) and it has been suggested that the high polyphenol content of the kauri foliage is responsible for the remarkable formation under old kauri trees (Bloomfield 1953; Van Breemen and Scott 1998) but this process has not been studied in much detail. It has further been suggested that kauri is better able to

14 General introduction

compete with angiosperm species on soils of low fertility (Halkett and Sale 1986; Burns and Leathwick 1996). In 2000, the Laboratory of Soil Science and Geology of Wageningen University, and the Nature Conservation and Plant Ecology Group of Wageningen University started the research program ‘Podzolisation under kauri, for better or worse?’. This program was supported by the Research Council for Earth and Life Sciences (ALW) with financial aid from the Netherlands Organisation for Scientific Research (NWO). The research project was split up in three parts, each covering one individual doctorate study. One study focused on the soil processes under kauri, giving details about the capabilities of kauri to create and about the formation of the organic layer. The second study, which is this present study, focused on the effects of kauri podzols on the vegetation, giving insight into the capability of plants to influence their own fitness via their effects on soil development. The third study focused on the distribution of soil types, the distribution of kauri trees and landscape dynamics, and resulted in the thesis ‘Modelling Landslide Dynamics in Forested Landscapes’ by Claessens (2005).

Kauri forest

Kauri forest can be regarded as a mosaic of vegetation patches dominated by kauri trees and patches dominated by other species (Bieleski 1979; Halkett and Sale 1986; Ogden and Stewart 1995; Burns and Leathwick 1996). These other species are often angiosperm tree species like tawa ( tawa), tarairi () or towai (Weinmannia silvicola), but the composition of these latter patches varies from region to region. On places where forest has disappeared, by for instance fire or a cyclone, patches of tea tree vegetation (Leptospermum scoparium and Kunzea ericoides) establish. It is there that many kauri seedlings can be found (Mirams 1957; Bieleski 1959). Kauri has small wind-dispersed and light-demanding seedlings. When tea tree vegetation becomes older, seedlings of kauri establish. Besides kauri seedlings also other tree seedlings, like for instance of Tanekaha (Phyllocladus trichomanoides) can be found. When enough kauri seedlings have established in the tea tree community, they will eventually replace the tea trees and become dominant (Esler and Astridge 1974; Pook 1978; Burns and Smale 1990). The kauri will form a forest of poles or ‘rickers’ and when the leading kauri poles are well above their associates, usually at 50 cm dbh, the conical ricker stem sheds its lower branches and develops an open, spreading crown above the developing canopy of the other species (Ogden and Stewart 1995). The estimated time to reach this stage is 120–170 years (Beveridge 1977).

15 Chapter 1

When the kauri shed their branches and spread their crown, a relatively even aged stand of mature kauri develops. Bieleski (1979) calls this stand ‘a mature kauri grove’. In this grove often a typical vegetation can be found with a herb and shrub layer dominated by species like kauri grass ( trinervia), cutty grass (Gahnia xanthocarpa), miniature tree fern (Blechnum fraseri) and soft mingimingi (Cyathodes fasciculata). Above this layer the forest is often more open than in other parts of the forest. In literature it is noted that kauri seedlings are rare under mature kauri trees, but they are present (Ogden et al. 1987). Also present under the kauri are, sometimes small and spindly versions of, trees that dominate

Study site Before the colonisation of New Zealand by Europeans there was about 1.5 million ha of indigenous forests containing kauri (Halkett 1983). Kauri has an even texture, is strong and durable and is excellent for sawing, machining and turning. These qualities made kauri the most important timber species in New Zealand for the first hundred years of European settlement (Ecroyd 1982). Extensive logging and clearance of kauri forest occurred, especially in the nineteenth and early twentieth centuries, and kauri forests were also burnt and cleared for agricultural use or when searching for ‘’ (Reed 1954; Halkett and Sale 1986). This kauri gum, the resin of kauri (Thomas 1969), was used to produce linoleum, varnish, wax matches, explosives and dentures (Ecroyd 1982). Only kauri trees which were too large to be milled and kauri stands in inaccessible locations were spared from logging and burning. By 1975 about 7000 ha untouched kauri forest remained. This forest is mainly confined to remote locations. On some of the cleared and burned areas communities of tea tree (Leptospermum scoparium and Kunzea ericoides) with regeneration of kauri appeared. The area of this second-growth forest exceeds 60000 ha (Halkett 1983, Ahmed and Ogden 1987). Most of the fieldwork of this thesis was carried out in the Waitakere Ranges, situated west of Auckland (Figure 1.2). Before logging and burning began in the th Waitakere Ranges, in the middle of the 19 century, the land was, except for the coastal fringe, ‘entirely covered with luxuriant forest‘ (Cheeseman 1972, in Esler 1983). Nowadays large parts of the vegetation of the Waitakere Ranges consist of ‘cut-over forest’, forest from which trees have been harvested. Another large area consists of ‘tea tree vegetation’, which established on cleared and burnt land. The main pockets of untouched forest in the Waitakere Ranges are situated in the Cascades Park and north of the lower Huia Reservoir (Esler 1983; Denyer et al. 1993). It is there that we conducted most of the fieldwork (Fig 1.2).

16 General introduction

other parts of the forest. When the initial cohort of kauri dies it is potentially replaced by a second cohort of kauri, which is less synchronous and less dense than the first (Ogden 1985). Many of the gaps created by the death of a mature kauri tree are requisitioned by faster growing angiosperm species (Ogden et al. 1987; Enright et al. 1993). Kauri becomes progressively restricted to sites where canopy gaps are large or frequent, like exposed ridge crests, and to sites of low fertility where young kauri trees compete most successfully (Ogden and Stewart 1995). Kauri is very well adapted for growth on sites of low fertility. Its mean leaf age is 2.5 years (Clarkson 1992), and the concentration of nitrogen, and potassium in kauri foliage is low compared with values of other species (Peterson 1962b). As a result of these characteristics kauri seedlings have a relatively high nutrient use efficiency (Peterson 1962a) and it is expected that kauri competes most successfully with angiosperm species on sites of low fertility (Pook 1979; Burns and Leathwick 1996). The low-fertility, high-disturbance sites can thus be regarded as the centre of the kauri niche and following larger catastrophes, kauri expands from these sites across the landscape (Ogden and Stewart 1995; Enright et al. 1999).

Figure 1.2 The Waitakere Ranges, situated west of Auckland at the Northern Island of New Zealand. The locations of the Huia area, Nihotupu area and Cascade area are indicated with 1, 2 and 3 respectively.

17 Chapter 1

Main hypothesis and outline of the thesis

The main hypothesis of this thesis is: kauri tannins can reduce nitrogen mineralisation in the soil below the crown of a kauri tree and thus create an environment where kauri seedlings are better able to compete with angiosperm seedlings, resulting in natural selection for kauri trees with high tannin concentrations. To test this hypothesis a number of studies was conducted, which I will present in the following chapters (Figure 1.3). In Chapter 2 I describe a study of the site conditions below and just outside the crown of five mature kauri trees. An important question was, if soil fertility is indeed lower below than outside the crown of kauri. In this study I also measured soil moisture availability and light conditions. Chapter 3 describes a study that was performed to test the hypothesis that the conditions below the kauri crown are favourable for kauri seedlings. I counted the number of tree and shrub seedlings below and outside the crown of the same five kauri trees as used in Chapter 2. By linking the data of the environmental variables with the seedling numbers, I investigated which site conditions are important for kauri seedlings and for its competitors. In Chapter 4 a field experiment is described that also focuses on the question: ‘which are the factors that affect the growth of kauri seedlings below the crown of kauri trees?’. I planted kauri seedlings and seedlings of an angiosperm tree species and applied different treatments to these seedlings. In the remaining chapters I focus on the role of tannins. Chapter 5 describes a laboratory study to test the hypothesis that kauri tannins can lower nitrogen mineralisation in the soil. Tannins were extracted from kauri leaves, and from the leaves of two other common New Zealand tree species. These tannins were added to soil organic material and this mixture was incubated in Petri-dishes. Afterwards the nitrogen mineralisation and the decomposition of the organic material were determined. In Chapter 6, three other studies on the role of tannins are presented. The first study also focuses on the effect of tannins on nitrogen mineralisation. In a litterbag experiment I compared the decomposition and mineralisation rate between kauri foliage with low and kauri foliage with high tannin concentrations. In another study I investigated if the tannin concentration of kauri foliage is high compared to the concentration in foliage of other New Zealand plant species. In a third study I tested if the variation in tannin concentration of kauri foliage is heritable, by comparing the tannin concentration in foliage of kauri seedlings. Finally, in Chapter 7 I integrate the results of the different studies.

18 General introduction

Is the tannin concentration higher in leaves of kauri than in leaves of other species? (6)

kauri

Outside the kauri Are site conditions different crown below and outside the crown of kauri trees? (2) Are seedling numbers different? (3) Why are seedling numbers different? (3, 4)

Are nitrogen availability and litter decomposition affected by tannins? (5, 6)

Figure 1.3. Outline of the thesis, with chapter numbers between brackets.

19 Chapter 2

20 Site conditions below kauri

Chapter 2

The effect of kauri trees on the availability of nutrients, water and light

Physical and chemical properties of soils are influenced by plants, and many of the soil properties influenced by plants feed back to the growth of the plants. Under the New Zealand kauri tree (Agathis australis) a thick organic layer is formed. We hypothesised that light intensity is higher and nutrient availability is lower under the kauri crown than just outside the kauri crown. We further hypothesised that these differences are related to the accumulation of soil organic matter below the kauri crown. To test these hypotheses we measured light availability, nutrient availability and soil moisture availability, in plots under and outside the crown of kauri trees. To investigate the consequences of differences in site conditions for the growth of kauri seedlings, we also measured the site conditions in vegetation of tea trees (Leptospermum scoparium and Kunzea ericoides) where kauri seedlings were abundant. We found that near the trunk of the kauri tree nutrient availability was lower than outside the crown. Further, the organic layer surrounding the kauri trunk was in some periods found to be a dry environment compared to the mineral soil further from the kauri trunk. The measurements suggest that conditions below the kauri crown are more similar to the conditions in tea tree vegetation than to the conditions in the plots just outside the kauri crown.

Based on: Verkaik, E., Braakhekke, W. G. The effect of kauri trees (Agathis australis) on the availability of nutrients, water and light. Submitted.

21 Chapter 2

Introduction

Physical and chemical properties of soils are influenced by plants, and many of the soil properties influenced by plants feed back to the growth of the plants (Berendse and Elberse 1990; Van Breemen 1993; Binkley 1995). A positive feedback might develop between plants and soil when plants produce litter that enhances their own competitive performance (Berendse 1994b; Van Breemen and Finzi 1998; Clark et al. 2005). Such positive feedbacks between plants and soil probably are important in the dynamics of natural ecosystems and need further study (Binkley and Giardina 1998; Van Breemen and Finzi 1998). Kauri (Agathis australis) forests in northern New Zealand are mixed forests of angiosperm and gymnosperm tree species. Under kauri trees thick layers of organic material develop (Silvester and Orchard 1999), which leads to increased weathering of the underlying 2 soil (Jongkind and Buurman 2006). In the organic layer up to 0.65 kg/m nitrogen may be contained, which apparently is inaccessible to plants (Silvester 2000). It has been suggested that kauri lowers nutrient availability in the soil (e.g. Bieleski 1959; Ecroyd 1982) and that kauri forest can be regarded a mosaic of site fertility (Ogden and Stewart 1995). On the organic layer below a kauri tree there is usually a dense herb layer dominated by species like kauri grass (Astelia trinervia) and cutty grass (Gahnia xanthocarpa). Above the herb layer the understorey under the crown is generally more open than in other parts of the forest (Bieleski 1979). In the present study we tested the hypotheses that: i) below mature kauri trees light intensity is higher than just outside the kauri crown; ii) below mature kauri trees nutrient availability is lower than just outside the kauri crown, and iii) the differences between soil conditions under and outside the kauri crown are related to the accumulation of soil organic matter under the kauri crown. We tested these hypotheses by comparing the conditions in plots under the crown of mature kauri trees (the “kauri plots”), with the conditions in plots just outside the crown of mature kauri trees (the “angiosperm plots”). Forest succession in northern New Zealand commonly starts with vegetation dominated by the two species commonly known as tea tree (Leptospermum scoparium and Kunzea ericoides, both Myrtaceae). For kauri seedlings, which need relatively open conditions (Pook 1979), this tea tree vegetation seems to provide optimal conditions for their growth (Mirams 1957; Ecroyd 1982). To evaluate the consequences of the differences in the site conditions below and outside the kauri crown for the growth of kauri seedlings, we also measured the site conditions in tea tree vegetation (the “tea tree plots”) where kauri seedlings were abundant. In all plots we measured soil water availability, light availability, and, as an indication of nutrient availability in the soil, the nutrient concentration in the leaves of hangehange

22 Site conditions below kauri

(Geniostoma rupestre var. ligustrifolium). This shrub species was common in each plot. To test the hypothesis that changes in conditions are related to thickness of the organic layer, we measured thickness of the organic layer and analysed the relation between thickness of the organic layer and the measured parameters.

Methods

Study areas All plots were situated in the Waitakere Ranges in New Zealand (36°57′S, 174°32′E) (Figure 2 1.2) and measured 100 m . We selected five mature kauri trees with diameters ranging from 1.1 to 2.3 m on different locations within the Waitakere Ranges, where there was no sign kauri had ever been logged. One tree was situated on a slope in the Nihotupu area, two trees were situated on the same ridge in the Huia area and two trees were situated on two different slopes in the Cascades area (Figure 1.2). Under the crown of each of these kauri trees one “kauri plot” was laid out randomly. In the vegetation dominated by angiosperm trees surrounding each kauri tree we selected one “angiosperm plot”. This plot was chosen to adjoin the edge of the kauri crown and was situated within 30 m distance from the kauri trunk. These angiosperm plots were selected to have the same inclination and orientation as the kauri plots and were not covered by the crown of a mature kauri tree. At one location, the angiosperm vegetation bordering the kauri tree did not have the same topography as the kauri tree. At that location, we situated the angiosperm plots nearby, on the same topography as the original kauri tree, and next to another mature kauri tree. Five “tea tree plots” were selected in secondary forest where kauri seedlings were abundant, one in the Nihotupu area, two in the Huia area and two in the Cascades area.

Nutrient concentration of hangehange leaves To compare the nutrient availability in the plots, we determined the nutrient concentration in the leaves of the shrub hangehange (Geniostoma rupestre var. ligustrifolium) which occurred in all plots. At the end of October 2004 in each plot three replicate samples were collected. For each sample leaves were collected from five to seven shrubs which had a height of 0.7-2.5 m. Only leaves with a length more than 4 cm were picked and per shrub three branches were sampled. From each branch we picked two young leaves, light green coloured on a yellowish petiole, and two older, dark green leaves. The leaves were air- and oven dried in New Zealand, weighted, transported to The Netherlands, where they were dried again (48 hours, 35 °C), ground and analysed. The ground material was digested with sulphuric acid, salicylic

23 Chapter 2

acid, hydrogen peroxide and selenium (Temminghoff et al. 2000) and the concentrations of nitrogen, phosphorus, potassium, calcium, magnesium and sodium in the digest were determined spectrophotometrically (Skalar, SAN plus system or AAS).

Soil moisture availability To compare soil moisture availability in the plots, we determined the actual water content (w), the soil bulk density (ρ) and the water content at pF 4.2 (wpF4.2), and calculated the available volumetric soil water content. In 2002 the water content was measured three times in the kauri and tea tree plots and in 2003 the water content was measured three times in all plots. For each measurement we took in each plot six soil samples at 0-7.5 cm depth and six soil samples at 7.5-15 cm depth using an auger. The depth was measured from the top of the litter layer and for each sample the thickness of the litter layer (L), the fermentation- plus humus layer (F+H), and the mineral layer was measured. In each sample the mass water content (w) was determined, by drying and weighing: mass of wet soil− mass of dry soil w = mass of dry soil The average soil bulk densities (ρ) were determined in each plot by taking six samples with an auger with known volume. The mass water content of the soil at permanent wilting point (wpF4.2) was determined with a pressure membrane apparatus (Pearcy et al. 1989). Near the kauri trunks a thick organic layer, dominated by humified material, is often present on top of mineral () material, while further from the trunk only a thin layer of undecomposed litter (L) is present on top of the clay. As a result some of the samples of the kauri plots were dominated by organic material and other samples were dominated by clay. The samples of the angiosperm and tea tree plots consisted mainly of mineral material (clay). We determined the wpF4.2 of three organic samples and three mineral samples from the kauri plots, of two mineral samples from the angiosperm plots, and of two mineral samples from the tea tree plots. The available water was calculated as:

availablevolumetric soil water content=−( w wpF 4.2 )i ρ. In the statistical analyses, for the samples from the kauri plots a distinction was made between samples dominated by organic material (F+H > 3.5 cm) and samples dominated by mineral (clay) material (F+H ≤ 3.5 cm).

Light measurements Canopy openness and total photon flux density in the plots were measured using a fish eye lens (Sigma 8 mm, f4). In May and June 2002 photos were taken in the kauri and tea tree

24 Site conditions below kauri

plots and in September 2004 photos were taken in the angiosperm plots. Under an evenly overcast sky in every plot on ten different points, evenly distributed over the plot, photos were taken 0.75 m above the ground with a camera (MZ-50 Pentax) that was levelled with a bubble level. The half-automatic mode of the camera was used with shutter speed set on 1/125, or slower under low light conditions, and with the use of the timed shutter release. Ilford ISO 125 black and white films were used, the films were scanned with a Nikon LS-2000 film scanner, the images were saved in BMP format and then converted (in Adobe Photoshop) to a grey scale image with a bit depth of eight bits. To calculate the canopy openness and the total photon flux density under the vegetation, the images were analysed with Winphot 5 (Ter Steege 1996). Using the manual threshold of the program, every image was converted twice to black and white pixels and the average value of the two threshold values was used (Englund et al. 2000). Nine days in the growing season were selected for the calculation of total photon flux density under the vegetation. Latitude was set at -36.5, longitude at -174.30, time zone at 12 and altitude at 100. Transmission of red, transmission of far red, the red far-red ratio, diffuse part and tau were set at the default values of 0.05, 0.45, 1.20, 0.15 and 0.60 respectively. For each day we have set the sun fraction at 50, used the standard overcast sky type and included diffuse canopy light.

Thickness of the organic layer and kauri diameter The thickness of the organic layer (litter layer plus fermentation layer plus humus layer) was measured in each plot using an iron rod with a diameter of 0.7 cm (Silvester and Orchard 1999). To estimate the thickness of the organic layer we probed on about 30 different positions in each plot, with the iron rod through the softer organic layer until the harder mineral layer was reached. We verified this method several times, by removing, by hand or with an auger, the organic layer from the mineral soil. The diameter of kauri trees (mature trees and ‘rickers’ or poles) in the plots was measured at 1.3m height.

Statistical analyses All statistical analyses were conducted with the SPSS statistical package for Windows (12.0.1). ANOVA with Tukey tests (significance level set at 0.05) were performed to test for differences in soil water availability, canopy openness and total photon flux density between the three plot types. Homogeneity of variance was tested before each ANOVA and when the variances were not equal Kruskal-Wallis tests were used. Differences in soil water availability between the kauri and angiosperm plots and between the samples from kauri plots dominated by organic material or clay were tested with a paired sample t-test. To test for differences in thickness of the litter layer a Kruskal-Wallis test was used since the variances of the different

25 Chapter 2

groups were not homogenous. Differences in nutrient concentrations between the kauri and angiosperm plots were tested with a paired non-parametric test (Wilcoxon) while other differences in nutrient concentrations were tested with a t-test. To evaluate the relation between thickness of the organic layer and other site conditions, Spearman correlations were calculated for the data of the kauri and angiosperm plots.

Results

Site conditions Nutrient concentrations in hangehange leaves differed between the angiosperm and kauri plots. Nitrogen, phosphorus and calcium concentrations were lower in the kauri plots (Figure 2.1). We found no significant differences between nutrient concentrations of hangehange leaves in the angiosperm and tea tree plots. Also, the nutrient concentrations of hangehange leaves in the kauri and tea tree plots were similar, except for the calcium concentration, which was higher in the tea tree plots than in the kauri plots (Figure 2.1). The weight of the leaves was similar for the kauri and angiosperm plots (data not shown). But the leaves collected in the tea tree plots weighted less than the leaves of the kauri plots, while the leaves of tea tree and angiosperm plots had similar weights (data not shown).

b a 1.6 a ab b a a b a a a 1.2 a kauri a 0.8 a a angiosperm tea tree 0.4 % of dry weight a b ab 0 N P K CaMgNa

Figure 2.1. Mean nutrient concentrations in hangehange leaves (% of dry weight) collected in the three plot types. Error bars indicate SE and different letters indicate significant differences (P < 0.05) between plot types.

The organic layer below the crown of the kauri trees was thicker than the organic layer outside the crown, in the angiosperm plots and in the tea tree plots (Table 2.1). The light conditions were not different between the three plot types (Table 2.1), despite some tea tree

26 Site conditions below kauri

and kauri plots appearing to be relatively open compared to the angiosperm plots. Also, for five of the six soil moisture measurements the available volumetric water content did not differ between the plot types. Only for the measurement in April 2002, at the depth 7.5-15 cm, the samples from the kauri plots dominated by organic material were dryer than the samples from the kauri plots dominated by clay and were dryer than the samples from the tea tree plots (Table 2.2).

Table 2.1. The mean (±SE) canopy openness (%), total photon flux density (mol/m2/day) and thickness of the organic layer (cm) in the three plot types. Column values followed by different letters are significantly different (P < 0.05).

Canopy openness Total photon flux density Thickness organic layer Kauri 10.2 ± 1.3a 3.59 ± 0.29a 10.6 ± 2.6a Angiosperm 8.2 ± 0.5a 3.20 ± 0.27a 1.5 ± 0.4b Tea tree 9.1 ± 0.9a 3.84 ± 0.56a 2.4 ± 1.1b

1,7

1,6

1,5

nitrogen content in hangehange leaves 1,4

0 5 10 15 20 thickness of the organic layer

Figure 2.2. Relationship between thickness of the organic layer (cm) and the nitrogen concentration in hangehange leaves (% of dry weight) in the kauri plots (●) and the angiosperm plots (x) (r = -0.794, p = 0.006).

27 Chapter 2

46.3a 29.8a 31.9a 26.2a 7.5-15 Leaf 0.224 0.091 0.430 1.000 -0.071 -0.103 -0.094 -0.030 May 2003 weight 0-7.5 23.3a 21.0a 27.4a 21.9a 26-5 to 31-5 2003 e correlations i Ca- the six soil moisture 0.513 1.000 -0.030 -0.612 -0.200 -0.382 0.685* -0.653* s followed by different letters 18.7a 23.0a 26.3a 20.9a 7.5-15 concentrat April 2003 P- i 0-7.5 13.3a 11.2a 18.0a 14.7a 28-4 to 1-5 2003 1.000 0.513 -0.350 -0.094 -0.650* 0.875** -0.863** -0.800** concentra 23.6a 27.0a 26.4a 21.9a 7.5-15 N- 1.000 -0.273 -0.103 0.685* -0.709* -0.692* 0.875** -0.794** March 2003 concentratio 0-7.5 15.2a 11.6a 15.6a 15.1a 26-3 to 3-4 2003 - 13.3a 23.2b 20.2b 7.5-15 0.459 0.472 1.000 Kauri -0.071 -0.692* -0.653* 0.912** diameter -0.800** April 2002 - 0-7.5 11.1a 11.0a 13.1a 16-4 to 18-4 2002 1.000 0.472 0.430 -0.382 0.661* 0.673* -0.709* -0.650* Canopy openness - 13.5a 21.4a 20.4a 7.5-15 n samples dominated by organic material or clay. Column value - 0.552 1.000 0.459 0.091 ation with other parameters were not significant. -0.273 -0.350 -0.200 0.673* February 2002 4-2 to 7-2 2002 0-7.5 13.1a 11.3a 15.9a flux density Total photon - 29.4a 32.9a 31.6a 7.5-15 0.224 1.000 0.552 -0.612 0.661* 0.912** -0.794** -0.863** Thickness of organic layer - January 2002 0-7.5 24.9a 29.9a 29.2a 13-1 to 21-1 2002

. Correlation coefficients between vegetation parameters and thickness of the organic layer (n=10). * P < 0.05, ** 0.01. Th The available volumetric soil moisture content (as % of volume) at two different depths (cm) in the three plot types for of K-concentration, Mg-concentration and Na-concentr Thickness of organic layer Total photon flux density Canopy openness Kauri diameter N-concentration P-concentration Ca-concentration Leaf weight Table 2.3 Date of measurement Depth of measurement Kauri-organic Kauri-clay Angiosperm Tea tree Table 2.2. measurements. For the kauri plots a distinction was made betwee are significantly different (P < 0.05).

28 Site conditions below kauri

▼

1-12-2003 . V 1-11-2003

1-10-2003 date, for the ▼▼ 1-9-2003 icated with 1-8-2003

1-7-2003 > 1-6-2003 > 1-5-2003 > 1-4-2003 1-3-2003 1-2-2003 ing the 20 days preceding a given 1-1-2003

1-12-2002 1-11-2002

1-10-2002

1-9-2002

1-8-2002

1-7-2002

1-6-2002 e Waitakere Ranges, received dur 1-5-2002 > 1-4-2002

> 1-3-2002 > 1-2-2002 1-1-2002

1-12-2001 1-11-2001

1-10-2001

1-9-2001

1-8-2001

1-7-2001

1-6-2001

1-5-2001 1-4-2001

1-3-2001 Rainfall (mm) at the Arataki Visitors Centre, situated in th 1-2-2001

1-1-2001 300 250 200 150 100 50 0 period 1-1-2001 to 31-12-2003. Rainfall data were provided by the Auckland Regional Council. Soil moisture measurements are ind Figure 2.3. precipitation sum

29 Chapter 2

Correlations Thickness of the organic layer and kauri diameter were positively correlated (Table 2.3) suggesting that organic material accumulates with age of the kauri tree. It appeared that thickness of the organic layer correlated positively with canopy openness too (Table 2.3). We did not find a correlation between thickness of the organic layer and the available volumetric soil water content for any of the measurements (data not shown). The results further suggest that plots with a thick organic layer had a lower nitrogen and phosphorus availability, since the nitrogen and phosphorus concentrations in the hangehange leaves correlated negatively with thickness of the organic layer (Table 2.3, Figure 2.2). The nitrogen and phosphorus concentrations also correlated negatively with canopy openness (Table 2.3). No correlations between light conditions and leaf weight were found, nor between leaf weight and nutrient concentration (Table 2.3).

Discussion

Light and nutrient availability below and outside the kauri crown We measured light conditions to test the first hypothesis, that light availability is higher below than just outside the kauri crown. In this study light was measured in autumn 2002 in the kauri and tea tree plots, and in spring 2004 in the angiosperm plots. The temperate forest in New Zealand is and light was not measured directly but was computed from grey- scale images of fish-eye photographs. All fish-eye photographs were taken on overcast days and therefore we think that comparison of the results of the two years is possible. The results of the light measurements (Table 2.1) do not support our first hypothesis. However, we observed a strong light intensity gradient from the darker edge of the crown to the more open area near the kauri trunk. We think that a comparison of the light conditions close to the kauri trunk, with the light conditions further from the trunk would have shown differences. The low nutrient concentrations which we found in hangehange leaves of shrubs situated below kauri crowns (Figure 2.1) support the second hypothesis which states that below mature kauri trees nutrient availability is lower than outside the kauri crown. An alternative explanation for the lower nutrient concentration in hangehange leaves below kauri than just outside the kauri crown, could be that there is a higher light availability below kauri than outside the kauri crown. More light might stimulate the growth of hangehange leaves thereby diluting the concentrations of nitrogen and phosphorus in these leaves. However, under field conditions nitrogen concentrations are often found to be higher in sun leaves than in shade leaves (Meletiou-Christou et al. 1994). Further, we found no correlation between the

30 Site conditions below kauri

weight of the hangehange leaves and canopy openness or photon flux density (Table 2.3). The weight of the leaves also did not correlate with any of the measured nutrient concentrations in these leaves (Table 2.3). Therefore we think that the lower nitrogen and phosphorus concentrations in the hangehange leaves below kauri than just outside the kauri crown do show that nutrient availability is lower under the crown. This finding confirms previous suggestions that kauri lowers nutrient availability in the soil (e.g. Bieleski 1959; Ecroyd 1982). Kauri forest can thus indeed be regarded a mosaic of site fertility, as suggested earlier (Ogden and Stewart 1995). Our soil moisture measurements show that in April 2002 the organic material under the kauri crown was dryer than the surrounding clay under the kauri crown (Table 2.2). Our first soil moisture measurement (in the third week of January 2002) was conducted after a rainy December and a rainy start of January (Figure 2.3) when the soil was saturated with water. At that time the available water content of the organic and the clay material under kauri were similar (Table 2.2), indicating that it is not the difference in material itself which caused the difference in water availability that we found later in the season. Kauri trees probably transpire more strongly than the surrounding vegetation since their crowns emerge above the main canopy where they are exposed to sun and wind. The fine roots of kauri are concentrated in the organic layer (Ecroyd 1982) and we propose that it is the uptake of water from the organic layer by the kauri tree which caused the dryness of the organic layer. The soil moisture measurement of April 2002 was preceded by a relative dry period (Figure 2.3), and because such dry periods are common in the Waitakere Ranges almost every year, we suspect that the organic layer is periodically a dry environment compared to the surrounding clay under the kauri. In April 2002 soil moisture availability was not measured in the angiosperm plots but data (Verkaik, unpublished data) suggest that during dry periods the organic layer under kauri can also be dryer than the surrounding clay soil in angiosperm vegetation outside the kauri crown.

The role of the organic layer The negative correlations between nutrient concentration in hangehange leaves and thickness of the organic layer and between canopy openness and thickness of the organic layer (Table 2.3, Figure 2.2) support the second hypothesis, that differences in conditions under and outside the crown are related to the organic layer under kauri. The positive correlation between canopy openness and thickness of the organic layer indicates that as organic material accumulates under the kauri tree, the vegetation on top of this organic layer becomes more open. The main explanation for this correlation seems to be the poor growth of many shrub and understorey tree species near the trunk of the kauri tree (Bieleski 1979). Our results

31 Chapter 2

indicate that this poor growth results from the lower nutrient availability under the kauri and the periodic dryness of the organic litter layer. The negative correlations between kauri diameter and the nutrient concentrations in hangehange leaves (Table 2.3) indicate that as the kauri tree grows older and thicker negative effects on nutrient supply increase.

Are conditions under kauri favourable for kauri seedlings? The measurements suggest that especially the organic layer under the kauri tree is a light and nutrient poor environment (Tables 2.1 and 2.3, Figures 2.1 and 2.2). In this regard the kauri plots are more similar to vegetation where kauri seedlings are abundant, in the tea tree plots, than to the angiosperm plots. A main difference between the kauri plots and the tea tree plots is that during dry periods the organic layer under kauri is probably dryer than the clay soil under tea tree (Table 2.2). We propose that this dryness is a main cause for the relative scarcity of kauri seedlings under mature kauri compared to the tea tree vegetation, since young kauri seedlings are susceptible to low soil moisture availability (Mirams 1957). In the long run periodically dryness of the soil may be an advantage to kauri, because - when they grow older- kauri seedlings can tolerate drought better than many other species (Bieleski 1959; Barton 1982; Stephens et al. 1999). The nutrient-poor conditions which we found under the kauri crown can be an advantage for kauri seedlings when they compete with angiosperm seedlings, since compared to other species kauri seedlings have a relative low nutrient requirement (Peterson 1962). On sites of higher soil fertility kauri seedlings are often outcompeted by faster growing angiosperm seedlings (Pook 1979; Bond 1989).

32 Site conditions below kauri

33 Chapter 3

34 Seedling distribution

Chapter 3

Seedling distribution below and outside the crown of kauri trees as affected by site conditions

It has often been suggested that plants can change soil characteristics via their litter to favour their own species. Because of its huge impact upon the soil the New Zealand kauri tree (Agathis australis) presents an interesting case for studying such a positive feedback between plant and soil. We hypothesised that: i) due to the poor soil conditions under kauri, seedlings of angiosperm tree species are rare under kauri compared to sites just outside the crown of kauri, and ii) due to the openness of the kauri crown kauri seedlings are relatively common under kauri trees compared to sites just outside the crown of kauri trees. We counted seedlings under and outside the crown of kauri trees and correlated the presence of these seedlings to measured site conditions. The results confirm the hypotheses and indicate that the establishment of kauri seedlings is favoured by the open canopy and high light intensities below kauri. The low nutrient availability under kauri appears to be unfavourable for the survival of angiosperm seedlings but not for the survival of kauri seedlings. Since the lower nitrogen availability under kauri seems due to the accumulation of nitrogen in the organic layer under kauri a positive feedback between kauri and the soil is likely.

Based on: Verkaik, E., Gardner, R. O., Braakhekke, W. G. Seedling distribution below and outside the crown of kauri trees (Agathis australis) as affected by site conditions. Submitted.

35 Chapter 3

Introduction

Many studies show that plant litter can change the characteristics of the soil that the plants are growing on (Van Breemen 1993; Berendse 1994a; Binkley 1995). It has often been suggested that these changes can be in favour of the species concerned. Studies show for instance how plant species which are adapted to poor soil conditions produce litter that decomposes slowly, thus keeping soil fertility at a low level and suppressing competing species that are adapted to less nutrient-poor substrates (Berendse 1998). It appears that a positive feedback might develop between plants and soil, whereby plants produce litter that further enhances their competitive dominance (Van Breemen and Finzi 1998). Binkley and Giardina (1998) suggested that a study of sites where the second generation of trees grows on sites influenced by the first generation would yield insights into the relative importance of such positive feedbacks for trees. Because of its huge impact on the soil the New Zealand kauri tree (Agathis australis) is an appropriate species to test these ideas. In another study we measured site conditions in plots situated below the crown of mature kauri trees and in plots just outside their crown (Chapter 2). The results of that study showed that nitrogen and phosphorus availability are lower under than outside the kauri crown. The results indicated that during dry periods, the organic layer under the crown can be drier than the surrounding clay soil under the kauri crown. It further appeared that as the kauri tree grows older and the organic layer gets thicker, the canopy of the vegetation becomes more open. Since kauri seedlings need high light intensities for their growth (Pook 1979) and probably can stand dry and nutrient poor conditions better than many angiosperm species (Bieleski 1959), the results of our other study suggested that conditions under the kauri crown are relatively favourable for the growth of kauri seedlings, and unfavourable for the growth of angiosperm seedlings in comparison to the conditions outside the kauri crown. In the present study we focussed on the distribution of seedlings and hypothesised that: i) due to the poor soil conditions under kauri, seedlings of angiosperm tree species are rare under kauri compared to sites just outside the crown of kauri, and ii) due to the openness of the kauri crown, kauri seedlings are relatively common under kauri trees compared to sites just outside the crown of kauri trees. To test these hypotheses, we counted seedlings of tree and shrub species in plots under the kauri crown and in plots outside the kauri crown. To investigate which site conditions are important for the growth of kauri seedlings, we also counted seedlings in plots in tea tree vegetation (Leptospermum scoparium and Kunzea ericoides, both Myrtaceae) where kauri seedlings were abundant. We analysed the relationships between seedling number and the site factors that we previously measured (Chapter 2).

36 Seedling distribution

Methods

Site conditions The plots were situated in the Waitakere Ranges in New Zealand (36°57′S, 174°32′E). A detailed description of the selection of the plots is given in Chapter 2. Five plots were chosen, each under a different mature kauri tree (the kauri plots). Five plots were situated nearby these plots, with the same inclination and orientation, but just outside the projection of the kauri crown, in vegetation dominated by angiosperm tree species (the angiosperm plots). To get an idea of the optimal conditions for the growth of kauri seedlings, five plots were situated in secondary forest where kauri seedlings were abundant (the tea tree plots). All plots were 100 2 m . A detailed description of the methods used to measure the site factors and of the results is given in Chapter 2. We determined the ‘plant available soil water’ by determining the actual volumetric water content of small soil samples and correcting for the water content at permanent wilting point. In 2002 available soil water was determined three times in the ‘kauri’ and ‘tea tree’ plots and in 2003 it was determined three times in all the plots. To compare nutrient availability in the plots we measured the nutrient concentrations of leaves of the common shrub hangehange (Geniostoma rupestre var. ligustrifolium). Light availability (photon flux density at ground level and canopy openness) in the plots was measured using a fish eye lens mounted on a camera. The thickness of the organic layer in the plots was measured by probing with an iron rod through the softer organic layer upon the harder clay layer.

Number of seedlings In each plot we counted the number of seedlings of trees and shrubs, and juvenile tree ferns. In January 2002 the seedlings were counted in the kauri and tea tree plots; in September 2004 they were counted in the angiosperm plots. We divided the seedlings in three height-classes; class 1 from 0 to 10 cm, class 2 from 10 to 30 cm and class 3 from 30 to 100 cm. Seedlings of shrubs were only counted as long as they were still mostly unbranched, tree ferns that were smaller than 1.0 m were also counted as ‘seedlings’, while seedlings of were not counted. Most seedlings could be identified, at least to genus level. We did not distinguish between Nestegis montana and N. lanceolata seedlings in the classes 1 and 2, although most were probably N. lanceolata seedlings since mature N. lanceolata plants were far more abundant than mature N. montana. Neither did we distinguish between the Coprosma species in class 1 and 2, except for the species C. lucida and C. grandifolia. Most of those

37 Chapter 3

undetermined Coprosma seedlings were probably C. arborea since this species was by far the most common in the vegetation. We did not distinguish between macrophyllus and M. ramiflorus seedlings. In some plots a small number (at most 5-10) of the smallest seedlings remained undetermined. We calculated the number of ‘non-gymnosperm seedlings’ by including seedlings of the following species: Beilschmiedia tarairi, , Coprosma arborea, Corynocarpus laevigatus, Cyathea dealbata, Dicksonia squarrosa, Dysoxylum spectabile, Elaeocarpus dentatus, Hedycarya arborea, Hoheria populnea var. populnea, Kunzea ericoides, Knightia excelsa, Litsea calicaris, Melicytus spp., Myrsine australis, Myrsine salicina, Nestegis lanceolata, Pittosporum tenuifolium, Pseudopanax arboreus, Pseudopanax crassifolius, Quintinia serrata, Rhopalostylis sapida, Schefflera digitata, Leptospermum scoparium and Vitex lucens. Since this group includes tree ferns as well as angiosperm seedlings, we will use the term ‘non-gymnosperm seedlings’ throughout the text.

Statistical analyses All statistical analyses were conducted with the SPSS statistical package for Windows (12.0.1). Differences in seedling numbers between kauri and angiosperm plots were tested with the paired non-parametric Wilcoxon test, and differences in seedling numbers between tea tree and kauri plots, and tea tree and angiosperm plots were tested with the non-parametric Mann-Whitney test. To evaluate the relations between the site factors and the presence or survival of the seedlings Spearman correlations were calculated for the data of all the plots, for the data of the combinations of two plot types (kauri and tea tree plots, kauri and angiosperm plots, and angiosperm and tea tree plots) and for the three individual plot types.

Results

Survival of seedlings Seedlings in the angiosperm plots were counted in September 2004 while seedlings in the kauri and tea tree plots were counted in January 2002. Therefore we did not compare absolute seedling numbers between the kauri and angiosperm plots, but compared the ratios of the seedlings in the different age classes, assuming a stable age class distribution. We found no difference between the kauri and angiosperm plots in the number of non-gymnosperm seedlings of class 2 expressed as percentage of class 1 (Figure 3.1). The number of non- gymnosperm seedlings of class 3 as percentage of class 1 was significantly lower under kauri than outside kauri (Figure 3.1), suggesting that under kauri a lower proportion of non-

38 Seedling distribution

gymnosperm seedlings survives than outside kauri, assuming a stable age class distribution. The numbers of kauri seedlings in classes 2 and 3 as percentage of kauri seedlings in class 1 were not different between the kauri and angiosperm plots (Figure 3.1). But in the angiosperm plots the number of kauri seedlings of class 3 as percentage of kauri seedlings of class 1 was lower than the number of non-gymnosperm seedlings of class 3 as percentage of class 1 (Figure 3.1). This suggests that in the angiosperm plots survival of kauri seedlings is lower than the survival of non-gymnosperm seedlings. In kauri plots these numbers were about equal (Figure 3.1), suggesting that survival of kauri seedlings and angiosperm seedlings there do not differ from each other.

60 AB 50 B 40 b

30 AB non-gym seedlings kauri seedlings 20 a A a 10 a % of seedlings in class 1 0 Class 2 Class 3 Class 2 Class 3

Kauri plots Angiosperm plots

Figure 3.1. Kauri seedlings (as % of kauri seedlings in class 1) and non-gymnosperm seedlings (as % of non-gymnosperm seedlings in class 1) for height classes two and three. Error bars indicate SE. Seedling numbers of the same seedling class but indicated with different letter are significantly different (P < 0.05).

Correlations between site conditions and seedling numbers Differences in nutrient availability appear to be important for non-gymnosperm seedlings, since the number of seedlings of class 3 correlated positively with the nitrogen concentration in hangehange leaves (Table 3.1). Also, the number of non-gymnosperm seedlings of class 2 as percentage of class 1, and of class 3 as percentage of class 1, correlated positively with the nitrogen concentration in hangehange leaves (Table 3.1, Figure 3.2). Further, the number of non-gymnosperm seedlings of class 2 and of class 3 correlated positively with the phosphorus

39 Chapter 3

concentration in hangehange leaves for the data of the kauri plots and tea tree plots together (data not shown). For the data of the kauri and tea tree plots together, the number of non- gymnosperm seedlings of class 2 as percentage of class 1 also correlated positively with the nitrogen concentration in the hangehange leaves (data not shown).

Table 3.1. Correlations between the site conditions and the absolute or relative (as percentage of seedlings in class 1) number of kauri and non-gymnosperm seedlings for the data of all the plots (n=15). * P < 0.05, ** P < 0.01.

Species Thickness Photon Canopy N P Kauri organic flux openness concentration concentratio diameter layer density hangehange n hangehange Abs. class1 Non-gym 0.171 0.132 0.143 -0.343 -0.081 0.383 Abs. class2 Non-gym -0.530* -0.191 -0.150 0.242 0.358 -0.307 Abs. class3 Non-gym -0.749** -0.291 -0.443 0.520* 0.496 -0.472 Class2 rel. class1 Non-gym -0.525* -0.257 -0.229 0.568* 0.373 -0.520* Class3 rel. class1 Non-gym -0.736** -0.282 -0.371 0.646** 0.433 -0.653** Abs. class1 Kauri 0.245 0.404 0.617* -0.410 -0.455 0.168 Abs. class2 Kauri -0.329 0.142 0.191 -0.091 -0.157 -0.304 Abs. class3 Kauri -0.271 0.261 0.033 -0.045 -0.287 -0.288 Class2 rel. class1 Kauri -0.560* -0.062 -0.069 0.135 0.028 -0.476 Class3 rel. class1 Kauri -0.380 0.250 -0.068 0.068 -0.218 -0.341

The number of non-gymnosperm seedlings of the classes 2 and 3 correlated negatively with thickness of the organic layer (Table 3.1). Also the number of non-gymnosperm seedlings in class 2 as percentage of class 1, and class 3 as percentage of class 1 correlated negatively with thickness of the organic layer (Table 3.1). These negative correlations between the absolute and relative number of larger non-gymnosperm seedlings and thickness of the organic layer were found too when the data of the angiosperm plots were left out of consideration (data not shown). The correlations suggest that in plots with a thick organic layer the survival of non- gymnosperm seedlings was lower than in plots with a thinner layer. The number of kauri seedlings of class 1 correlated positively with canopy openness for the data of all the plots (Table 3.1), and also for the data of the kauri plots and tea tree plots together (data not shown). This indicates that small kauri seedlings are favoured by open conditions. In plots with a thick organic layer the survival of small kauri seedlings is low, indicated by the negative correlations between thickness of the organic layer and the number

40 Seedling distribution

of kauri seedlings of class 2 as percentage of class 1 (Table 3.1). This negative correlation was also found for the data of the kauri and tea tree plots together (data not shown). The density of kauri seedlings did not correlate with nitrogen concentration or phosphorus concentration in hangehange leaves. Sites with more available soil moisture appear to have less small non-gymnosperm seedlings, since we found negative correlations between the number of non-gymnosperm seedlings of class 1 and soil moisture availability in the upper soil layer for the measurements of March 2003 and May 2003 (Table 3.2, data of the kauri and the angiosperm plots together). Negative correlations between soil moisture availability and the number of non-gymnosperm seedlings of class 1 were also found for the data of the kauri plots alone and for the data of the angiosperm plots alone (data not shown). The survival of small angiosperm seedlings seems related to soil moisture availability since we found a positive correlation between soil moisture availability and the number of non-gymnosperm seedlings of class 2 as percentage of class 1 (Table 3.2). This positive correlation was also found for the data of the tea tree and kauri plots together and of for the data of the kauri plots alone (data not shown). For larger non-gymnosperm seedlings we did not find correlations with soil moisture availability.

Table 3.2. Correlations between soil moisture availability, for the three soil moisture measurements in 2003 at the two different depths, and the absolute or relative (as % of seedlings in class 1) number of kauri seedlings and non-gymnosperm seedlings, for the data of the kauri and angiosperm plots (n=10). * P < 0.05 level, ** P < 0.01.

March 2003 April 2003 May 2003 0-7.5 cm 7.5-15 cm 0-7.5 cm 7.5-15 cm 0-7.5 cm 7.5-15 cm Abs. class1 Non-gym -0.636* -0.127 -0.576 -0.079 -0.867** -0.236 Abs. class2 Non-gym -0.067 0.097 -0.219 -0.298 -0.426 -0.365 Abs. class3 Non-gym 0.127 0.042 -0.079 -0.285 -0.006 -0.212 Class2 rel. class1 Non-gym 0.636* 0.382 0.455 0.079 0.552 0.067 Class3 rel. class1 Non-gym 0.467 0.067 0.345 -0.139 0.539 -0.042 Abs. class1 Kauri 0.225 0.067 0.231 0.201 0.316 0.316 Abs. class2 Kauri 0.701* 0.785** 0.474 0.454 0.259 0.629 Abs. class3 Kauri 0.415 0.182 -0.061 -0.216 0.242 0.216 Class2 rel. class1 Kauri 0.808** 0.925** 0.614 0.653* 0.356 0.743* Class3 rel. class1 Kauri 0.415 0.182 -0.061 -0.216 0.242 0.216

The survival of small kauri seedlings appears related to soil moisture availability too, since we found positive correlations between soil moisture availability for the measurement of March 2003 and the number of kauri seedlings of class 2 (Table 3.2). Further, we found

41 Chapter 3

positive correlations between the number of kauri seedlings in class 2 as percentage of class 1 for the measurements of March 2003 in both soil layers, and in April 2003 and May 2003 in the deeper soil layer (Table 3.2). These positive correlations were also found for the data of the kauri plots alone and for the data of the angiosperm plots alone (not shown). Like for larger non-gymnosperm seedlings we found no correlation between larger kauri seedlings and soil moisture availability.

70

60

50

40

30

20

10

non-gymnosperm seedlings in class 3 (as % of 1) 0

1,20 1,30 1,40 1,50 1,60 1,70 nitrogen concentration in hangehange leaves

Figure 3.2. Relationship between nitrogen concentration in hangehange leaves (% of dry weight) and the number of non-gymnosperm seedlings of height class 3 as percentage of non-gymnosperm seedlings in height class 1 (r = 0.646, p = 0.009), for the kauri plots (●), the angiosperm plots (x) and the tea tree plots (▼).

Discussion

Angiosperm seedlings The observation that the number of non-gymnosperm seedlings of class 3 as percentage of class 1 is lower under than outside kauri (Figure 3.1) supports the hypothesis that seedlings of angiosperm species are rare under kauri. This result might be affected by the different years that the seedlings were counted in the two plot types. However, also the correlation tests

42 Seedling distribution

suggest that conditions under kauri are not favourable for the growth and survival of angiosperm seedlings. Both the absolute number of non-gymnosperm seedlings of class 3 and the relative number of non-gymnosperm seedlings of class 2 as percentage of class 1 and of class 3 as percentage of class 1, correlated negatively with thickness of the organic layer (Table 3.1). These correlations were also found when the angiosperm plots were left out of consideration. That result is not affected by the seedlings having been counted in different years since the seedlings in the kauri and tea tree plots were counted in the same year. An explanation for the negative correlation between survival of non-gymnosperm seedlings and thickness of the organic layer seems to be the low availability of nutrients in kauri plots. Nutrient availability is lower under kauri than outside the crown of kauri and correlates negatively with thickness of the organic layer (see Chapter 2). The positive correlations that we found between nitrogen availability and the presence or survival of non-gymnosperm seedlings (Figure 3.2, Table 3.1), and between phosphorus availability and the presence of non-gymnosperm seedlings (data not shown), indicate that the presence of these seedlings is hampered by nutrient deficiency. Significant correlations with nutrient availability were again also found when the data of the angiosperm plots were omitted.

Kauri seedlings Our hypothesis that conditions on the litter layer under kauri are favourable for kauri seedlings is supported by part of the results. The number of kauri seedlings of height class 1 correlated positively to canopy openness (Table 3.1), and this positive correlation was also found when the data of the angiosperm plots were left out of consideration. There are other studies suggesting that high light intensities favour the growth of kauri seedlings (Bieleski 1959; Pook 1979; Ogden et al. 1987; Wright 1993). Since canopy openness correlated positively to thickness of the organic layer (Chapter 2) the density of kauri seedlings is expected to be positively correlated to thickness of the organic layer. For the kauri plots we did find a positive correlation between the number of kauri seedlings of height class 1 and thickness of the organic layer (data not shown). The positive correlations between soil moisture availability and the survival of small kauri seedlings (Table 3.2) indicate that small kauri seedlings in kauri forest are susceptible to drought (Mirams 1957). Since in kauri plots and tea tree plots nutrient and light availability are alike and soil moisture availability differs (Chapter 2), we think that drought is responsible for the lower survival of small kauri seedlings in the kauri plots as compared to the tea tree plots. Larger kauri seedlings might be less susceptible to drought, since we found no correlation between soil moisture and the presence of larger kauri seedlings. Other studies have shown that established kauri seedlings are drought resistant and have a high nutrient use

43 Chapter 3

efficiency compared to other species (Bieleski 1959; Peterson 1962; Barton 1982; Stephens et al. 1999). The presence of several hundreds of established kauri seedlings and saplings per hectare in mature kauri forest (Ogden et al. 1987) agrees with these findings.

A positive feedback between kauri and the soil? The correlations we found suggest that conditions below the kauri crown are favourable for the initial phase of seedling establishment. The relatively open canopy above the organic layer under kauri favours the establishment of kauri seedlings, and the periodically dry conditions of the organic layer under kauri might favour the establishment of non-gymnosperm seedlings. However, in a later stage many of those kauri and non-gymnosperm seedlings do not survive under kauri. In the kauri plots the survival percentages of kauri seedlings and of non-gymnosperm seedlings were low but did not differ significantly from each other. In the angiosperm plots however the survival percentage of kauri seedlings was lower than the percentage of non-gymnosperm seedlings. We suspect that outside the kauri crown, where nutrient availability is somewhat higher, kauri seedlings lose the competition for light from faster growing angiosperm seedlings, as suggested for gymnosperm species in general (Bond 1989) and for kauri specifically (Pook 1979; Burns and Leathwick 1996). Under kauri, the kauri seedlings appear better able to compete with angiosperm seedlings since here the latter seem hampered in their growth by lower nutrient availability. Since kauri produces litter that decomposes slowly, leading to the sequestration of nitrogen in the organic layer under kauri (Silvester 2000; Chapter 2) a positive feedback between kauri and the soil is likely.

44 Seedling distribution

45 Chapter 4

46 Growth of seedlings

Chapter 4

The decrease of the soil resource by New Zealand kauri trees increases the relative fitness of their seedlings

Tree species can affect the soil they are growing on and this might influence their fitness. The New Zealand tree species kauri (Agathis australis (D. Don) Lindl.), which grows in mixed angiosperm-gymnosperm forests, has a substantial effect upon the soil. We studied the hypotheses that: 1) low soil moisture availability below mature kauri trees hampers growth of kauri seedlings and angiosperm seedlings, 2) low nutrient availability below kauri trees hampers only angiosperm seedlings, and 3) angiosperm seedlings are hampered more than kauri seedlings by the conditions below kauri trees. We tested these hypotheses by planting seedlings of kauri and mapau (Myrsine australis (A. Rich) Allan) under kauri trees and applying the following treatments: removal of herbs, removal of litter, removal of nutrient limitation by fertilization, and elimination of root competition of mature kauri trees by trenching. The results indicate that low soil moisture availability, or the combination of low soil moisture availability and low nutrient fertility, hampers the growth of kauri as well as mapau seedlings below kauri trees. The mapau seedlings are hampered relatively more than the kauri seedlings which might result in an increased relative fitness of the latter.

Based on: Verkaik, E., Berendse, F., Gardner, R.O. The decrease of the soil resource by New Zealand kauri (Agathis australis (D. Don) Lindl.) trees increases the relative fitness of their seedlings. Submitted.

47 Chapter 4

Introduction

Trees have important impacts on soil processes such as nutrient mineralisation and organic matter dynamics. Different tree species can have different effects which might even lead to differences in the understory vegetation between tree species (Binkley 1995; Hommel et al. 2002). The effects trees have upon the soil can also affect the fitness of the individual trees themselves. Frelich et al. (1993) studied a forest mosaic of patches dominated by hemlock (Tsuga canadensis) or sugar maple (Acer saccharum) in Upper Michigan (USA). This forest mosaic was found to result from the influence these species have, via the soil, upon each other’s seedlings. It is thought that the roots of the hemlock seedlings cannot penetrate the thick coarse litter layer that is present under sugar maple while sugar maple seedlings are thought to be intolerant to the low nitrogen availability under hemlock. Reviewing the literature, Binkley and Giardina (1998) found examples of tree species that have a positive effect upon their own fitness, and also examples of tree species that have a negative effect upon their own fitness due to their effects on soil processes. To get more insight in the relative importance of the feedbacks between trees and soil they suggested studying the second generation of trees on sites influenced by the first generation. An interesting species for such a study is the New Zealand kauri tree (Agathis australis (D. Don) Lindl.) because it has a huge impact upon the soil. Kauri is a large, emergent evergreen gymnosperm tree that grows in mixed angiosperm-gymnosperm forests in northern New Zealand. Under its crown a thick organic layer develops. The large amount of nitrogen in this layer seems to be inaccessible to plants (Silvester and Orchard 1999; Silvester 2000). It appears that the organic layer leads to increased weathering of the underlying soil (Swindale 1957; Jongkind and Buurman 2006). A herb layer under the crown of a kauri is typically dominated by kauri grass (Astelia trinervia) and cutty grass (Gahnia xanthocarpa). Seedlings of kauri are present below mature kauri trees (Ogden et al. 1987) but are relatively scarce. Despite several studies (e.g. Mc Kinnon 1945; Mirams 1957; Bieleski 1959) the cause of this scarcity is still unclear (Ecroyd 1982). A previous study (Chapter 2) suggests that nutrient availability is lower below the crown of kauri trees than outside the crown of the kauri trees and indicated that during dry periods, the organic layer under the crown can be drier than the surrounding clay. The presence of angiosperm seedlings in kauri forest correlated negatively with drought as well as with low nutrient availability, while the presence of kauri seedlings correlated negatively only with drought (Chapter 3). In the present study we tested three hypotheses: 1) low soil moisture availability below kauri hampers growth of kauri seedlings and angiosperm seedlings, 2) low nutrient

48 Growth of seedlings

availability below kauri only hampers angiosperm seedlings, and 3) angiosperm seedlings are hampered more than kauri seedlings by the conditions below kauri. In this study mapau (Myrsine australis (A. Rich) Allan), a common understory tree in kauri forests (Wardle 1991), was used as an example of an angiosperm species. We planted seedlings of kauri and of mapau below mature kauri trees and applied the following treatments: the removal of herbs (‘no herbs’); the removal of herbs and elimination of root competition (‘trenching’); the removal of herbs and the addition of slow-release fertilizer (‘fertilization’), the removal of herbs and all coarse litter (‘no litter’) and a control (‘control’). The experiment was run for 30 months and light availability, soil moisture availability and, as an indication of nutrient availability, the nutrient concentrations of the leaves of the seedlings, were measured.

Methods

Design of the experiment The field experiment was started in May 2002 in the Waitakere Ranges (36°57′S, 174°32′E) (Figure 1.2) below five mature kauri trees. Two kauri trees were situated in the Cascade area, on two different slopes; two were situated in the Huia area, both on the same ridge; and one tree was situated in the Nihotupu area, on a slope. Below the crown of each kauri tree five 2 plots of 1m were selected. These plots were situated where no trees, shrubs or tall herbs were growing, and were situated on average 5-10 m from the trunks of the kauri trees. The fermentation plus humus layer (F+H) in the plots was thin (0-3 cm) to moderately thick (10 cm), and the coarse litter layer (L) was on average 0.5-3cm thick. The treatments were randomly assigned to the five plots at each site. In each plot we planted six one year old kauri seedlings and six two year old mapau seedlings which were obtained from a nursery (Oratia Native Plant Nursery, Auckland). The seedlings were about 20 cm high, and roots of kauri seedlings were 7-15 cm long and roots of mapau seedlings were 7-10 cm long. Prior to planting, all herbs and seedlings were removed from all plots except for the ‘control’ plots. For the ‘no herbs’ plots this weeding was the only treatment. Around each ‘trenching’ plot a trench of 40 cm depth was dug, a plastic sheet was placed in the trench to prevent roots growing back in the plot and the trench was filled with soil. In the ‘fertilization’ plots one granule slow-release fertilizer (‘Osmocote Plus for Pots and Garden’), containing on average 3.8 mg nitrogen, 1.1 mg phosphorus and 2.5 mg potassium, was put into the soil next to each seedling. From the ‘no litter’ plots we removed the L-layer, consisting of the coarse undecomposed litter. Once every 3.5 months the herbs and seedlings were removed from all plots except for the control plots, and once every 3.5

49 Chapter 4

months one granule slow-release fertilizer was added to the seedlings in the plots with the ‘fertilization’ treatment, and coarse litter was removed from the plots with the ‘no litter’ treatment.

Measurements Canopy openness and total photon flux density in the plots were measured in October 2004 using a fish eye lens (Sigma 8 mm, f4). At two different positions in each plot, a photograph was taken 1.0 m above the ground with a camera (MZ-50 Pentax) which was levelled with a bubble level. The half-automatic mode of the camera was used with shutter speed set on 1/125, or slower under low light conditions, and with the use of the timed shutter release. Ilford ISO 125 black and white films were used, the films were scanned with a Nikon LS- 2000 film scanner, the images were saved in BMP format and then converted (in Adobe Photoshop) to a grey scale image with a bit depth of eight bits. To calculate canopy openness and the total photon flux density under the vegetation, the images were analysed with Winphot 5 (Ter Steege 1996). Using the manual threshold of the program, every image was converted twice to black and white pixels and the average value of the two threshold values was used (Englund et al. 2000). Nine days in the growing season were selected for the calculation of total photon flux density under the vegetation. Latitude was set at -36.5, longitude at -174.30, time zone at 12 and altitude at 100. Transmission of red, transmission of far-red, red far-red ratio, diffuse part and tau were set at their default values of 0.05, 0.45, 1.20, 0.15 and 0.60 respectively. For each day we have set the sun fraction at 50, used the standard overcast sky type and included diffuse canopy light. Soil moisture availability was determined in autumn 2003 (the last week of March and the first week of April) and in November 2004 when the experiment was ended. To compare soil moisture availability, we determined the actual water content (w), the soil bulk density (ρ) and the water content at pF 4.2 (wpF4.2), and calculated the available volumetric soil water content. At two different positions in each plot soil samples were taken with an auger at two depths (0-7.5 cm depth and 7.5-15 cm depth). The depth was measured from the top of the litter layer and for each sample the thickness of the litter layer (L), the fermentation- plus humus layer (F+H), and the mineral layer was measured. In each sample the mass water mass of wet soil− mass of dry soil content (w) was determined, by drying and weighing: w = mass of dry soil The average soil bulk densities (ρ) were determined under each kauri by taking six samples with an auger with known volume. The mass water content of the soil at permanent wilting point (wpF4.2) was determined with a pressure membrane apparatus (Pearcy et al. 1989). We determined the wpF4.2 of in total three samples of the organic layer (mainly consisting of

50 Growth of seedlings

humus material) and of three samples of the mineral layer. The available water was calculated as: availablevolumetric soil water content=−( w wpF 4.2 )i ρ. At the start of the experiment, during the experiment with 3.5 months interval, and at the end of the experiment, the height of all seedlings was measured. At the end of the experiment the living seedlings were cut off and dried (95 hours at 25°C) and per seedling the total above-ground biomass and the diameter of the root collar was measured. To measure nutrient concentrations of the leaves of the seedlings, for each plot a sample consisting of 160 kauri leaves and 120 mapau leaves was collected. The samples were transported to the Netherlands where the leaves were dried (48 hours, 35 °C), ground and analysed. The ground material was digested with sulphuric acid, salicylic acid, hydrogen peroxide and selenium (Temminghoff et al. 2000) and the concentrations of nitrogen, phosphorus, potassium, calcium, magnesium and sodium in the digest were determined spectrophotometrically (Skalar, SAN plus system or AAS).

Statistical analyses All statistical analyses were conducted with the SPSS statistical package for Windows (12.0.1). Before analyses, above-ground biomass was log-transformed to achieve homogeneous variances, and mean height growth of the seedlings was calculated for the surviving seedlings. Differences between treatments in canopy openness, light availability, soil moisture availability, amount of nutrients in leaves of seedlings, nutrient concentrations of leaves of seedlings, height of seedlings, height growth of seedlings, above-ground biomass of seedlings and diameter of root collar of seedlings were tested with two-way ANOVA’s and Tukey tests (significance level set at 0.05), with ‘treatment’ as fixed and ‘site’ as random factor and no interaction. Before running each two-way ANOVA, homogeneity of variation was tested with the Levene’s test and ‘treatment’ as only fixed factor. When the variation was not homogenous across treatments the non-parametric Friedman test was used. Differences between treatments in the survival of seedlings were also tested with the Friedman test. For each treatment differences between mapau and kauri in the reduction of height, above-ground biomass and diameter of root collar of the control, compared to the height, above-ground biomass and diameter of root collar of the treatments, were tested with a with a two-way ANOVA with ‘species’ as fixed and ‘site’ as random factor and no interaction. Before running this two-way ANOVA, homogeneity of variation was tested with Levene’s test and ‘species’ as only fixed factor. When variation was not equal across species the non-parametric Wilcoxon test was used. This test was also used to compare the survival between the two species.

51 Chapter 4

Results

In autumn 2003 soil moisture availability was significantly higher in the trenched plots than in the other plots (Table 4.1). At the end of the experiment nitrogen concentrations of the kauri and mapau leaves and the total amounts of nitrogen in the leaves were higher in the trenched plots than in the other plots (Table 4.2), indicating that trenching also affected nutrient availability in the plots. The phosphorus concentration of the seedlings was similar for all plots but at the end of the experiment the total amount of phosphorus in the mapau leaves was higher in the trenched plots than in the other plots (Table 4.2). Canopy openness and photon flux density were similar in all treatments (Table 4.1).

2 Table 4.1. Mean canopy openness (%), photon flux density (mol/m /day) and soil moisture availability (% of soil volume) in the field experiment in the different treatments. Row values followed by different letters are

significantly different (P < 0.05).

Horizon Control No herbs Trenching Fertilization No litter Canopy openness 13.5a 11.7a 13.3a 12.7a 11.9a Photon flux density 5.0a 4.1a 5.6a 4.2a 4.3a Soil moisture, autumn 2003 Upper 8.21a 11.29a 22.09b 11.74a 12.40a Soil moisture, autumn 2003 Lower 24.22ab 21.85a 35.41b 24.59ab 23.08a Soil moisture, spring 2004 Upper 15.39a 20.49a 33.73a 24.86a 15.13a Soil moisture, spring 2004 Lower 36.91a 32.10a 48.03a 33.71a 27.92a

The kauri seedlings in the trenched plots had a significantly larger height increment than the kauri seedlings in the control plots (Table 4.3). The mapau seedlings in the trenched plots had a larger height increment, had a much larger above-ground biomass and had a bigger root diameter than the seedlings in the control plots (Table 4.3). We found no significant differences in the survival of the kauri seedlings between the treatments and also the survival of the mapau seedlings was similar in all treatments (Table 4.3). The survival of mapau seedlings was high at all the five sites while the survival of kauri seedlings was on average 75% at four sites and only 27% at one site. Near this site, dying and dead older rickers and mature kauri trees were present, and the cortex of the root collar of the kauri seedlings which died at this site could often readily be stripped from the stele. This symptom is often found in plants which are infected by the root rot cinnamomi, which has been reported for the Waitakere Ranges (Podger and Newhook 1971).

52 Growth of seedlings

Table 4.2. Mean nitrogen and mean phosphorus concentration (% of leaf dry weight) and the mean amounts of nitrogen and phosphorus in the leaves of the seedlings (mg/plant) at the end of the experiment. Row values followed by different letters are significantly different (P < 0.05).

Control No herbs Trenching Fertilization No litter Kauri N concentration 0.56a 0.58a 0.97b 0.59a 0.60a P concentration 0.06a 0.08a 0.07a 0.06a 0.07a Amount of N 6.8a 7.5a 15.3b 6.3a 5.7a Amount of P 0.8a 1.0a 1.2a 0.7a 0.7a Mapau N concentration 0.85a 0.86a 1.07b 0.82a 0.77a P concentration 0.06a 0.06a 0.06a 0.05a 0.06a Amount of N 9.4a 8.3a 57.9b 13.6a 6.7a Amount of P 0.7a 0.6a 3.2b 0.8a 0.5a

To test the hypothesis that angiosperm seedlings are hampered more than kauri seedlings by the conditions below kauri we compared the response of the two species to the treatments. In none of the treatments there was a difference between the survival of kauri and mapau (data not shown) but growth differed between the two species. Mean height growth of the kauri seedlings was -0.7 cm in the control plots while the mapau seedlings grew on average 7.6 cm (Table 4.3). As a result the height at the end of the experiment, expressed as percentage of initial height, was significantly higher for mapau than for kauri (Figure 4.1).

Table 4.3. The mean survival (% of the seedlings at the start of the experiment), height growth (cm), above-ground biomass (g) and diameter of root collar (cm) of the mapau and kauri seedlings in the different treatments. Row values followed by different letters are significantly different (P < 0.05).

Control No herbs Trenching Fertilization No litter Kauri Survival 63.3a 73.3a 70.0a 53.3a 66.7a Height growth -0.7a 0.6ab 4.2b -0.7a 0.8ab Above-ground biomass 2.72a 3.10a 4.02a 2.74a 2.63a Root diameter 0.38a 0.43a 0.44a 0.41a 0.40a Mapau Survival 96.7a 93.3a 86.7a 100.0a 83.3a Height growth 7.6a 10.1a 24.1b 10.2a 5.6a Above-ground biomass 2.98a 2.76a 13.92b 4.19a 2.63a Root diameter 0.42a 0.44a 0.68b 0.48a 0.45a

53 Chapter 4

300 kauri 250 * mapau

200 * * 150

100

50

0

r ol n tr ing io n h t co iza o litte

height of seedlings (% of initial height) initial of seedlings (% height il n no herbs trenc rt fe

Figure 4.1. Mean height of the kauri and mapau seedlings at the end of the experiment (as % of initial height) for the treatments. Error bars indicate SE. The mean height differed significantly (P < 0.05) between the kauri and mapau seedlings in the control, the trenching treatment and the fertilization treatment (marked *).

To analyse the negative effects on the seedlings of the factors which were released by the treatments, we expressed the reduction of the mean height, root diameter and above-ground biomass of the control at each site as percentages of the values measured in the treatments. Except for the trenching treatment for both species the reduction of the growth in the control plots was about 0% compared to the growth in the treatments (Figure 4.2, data of height growth and root core diameter are not shown). This result suggests that, for both species, the presence of the herb layer, nutrient deficiency or the presence of a litter layer did not reduce the growth of the seedlings in the control plots. For the trenching treatment the reductions were clearly higher than 0% and the reductions of above-ground biomass and root-core diameter were significantly higher for mapau than for kauri (Figure 4.2, data of height growth and root core diameter are not shown). For both species, competition by tree roots reduced the growth of the seedlings in the control plots, and mapau seedlings suffered much more than kauri seedlings.

54 Growth of seedlings

100 * Kauri 80 Mapau

60

40

20

0

-20 reduction of biomass (%)

-40

-60 No herbs Trenching Fertilization No litter

Figure 4.2. The reduction of the above-ground biomass of the control compared to the above-ground biomass of the treatments (calculated as: reduction of biomass = above− ground biomass control 100−  100* ). Error bars indicate SE. The above− ground biomasstreatment reduction of the biomass differed significantly (P < 0.05) between the kauri and mapau seedlings for the trenching treatment (marked *).

Discussion

Effects of treatment on growth of seedlings The trenching treatment was the only treatment which caused changes in soil water and nutrient availability compared to the control (Tables 4.1 and 4.2), resulting in a more rapid growth of seedlings (Table 4.3). The trenching treatment consisted of the removal of herbs and seedlings and the digging of a trench. Since the removal of herbs and seedlings in the ‘no herbs’ treatment neither influenced nutrient and water availability (Tables 4.1 and 4.2) nor affected the growth of seedlings (Table 4.3) it must have been the trenching itself that caused the effects in the trenched plots. During the digging of the trenches, especially roots of kauri were cut, and the results thus suggest that excluding the roots of the mature kauri trees increases nutrient and water availability below kauri trees leading to an increased growth of seedlings.

55 Chapter 4

Increasing nutrient availability by adding fertilizer had no significant effect on the growth of the kauri seedlings, as indicated by the results of the fertilization treatment (Table 4.3). Peterson (1962a; 1962b) studied the nutrient concentrations and nutrient deficiency symptoms in kauri seedlings. Comparison of Peterson’s data with the nutrient concentrations in the leaves of our seedlings (Table 4.2) indicates that at the end of the experiment the kauri seedlings in the fertilized plots were nitrogen - and phosphorus - deficient. In the fertilization treatment, every 3.5 months one granule of slow-release fertilizer, containing on average 3.8 mg nitrogen and 1.1 mg phosphorus, was added to each seedlings. Since at the end of the experiment the leaves of the control kauri seedlings contained only about 7 mg nitrogen and 0.8 mg phosphorus, the seedlings in the fertilized plots were provided each 3.5 months with a significant amount of nutrients. Even so, nutrient concentration and growth of kauri seedlings were not altered by the extra nutrients compared to the control plots. For the mapau seedlings, the amounts of nitrogen in the leaves and the above-ground biomass were slightly, but not significantly, larger at the end of the experiment for the fertilization treatment than for the control (Tables 4.2 and 4.3). The results of the fertilization treatment thus indicate that providing extra nutrients does not in itself result in an increased growth or nutrient uptake of the kauri and mapau seedlings. In contrast, the extra nutrients which were available in the trenching treatment were taken up by the kauri and the mapau seedlings. This difference between treatments suggests that low soil moisture availability in the fertilization treatment limited the uptake of the added nutrients. With respect to the first and second hypothesis the results suggest that the main cause of the poor growth of young kauri and mapau seedlings below kauri is low soil moisture availability or the combination of low soil moisture availability and low nutrient availability. Low nutrient availability in itself appears not to be the cause for the poor growth of kauri and mapau seedlings. Also, root or light competition by herbs, or the presence of coarse litter does not seem to hamper the growth of the seedlings.

Consequences for competition between kauri and mapau seedlings During the experiment the kauri seedlings in the control plots had a slight negative growth while the mapau seedlings grew almost 8 cm (Table 4.3, Figure 4.1). Kauri seedlings are known to be able to persist when growing conditions are not favourable, as in shaded situations (Ogden et al. 1987). The negative growth for kauri in our experiment can be explained by a dieback of the main, leading shoot of some seedlings during the experiment. Further, some seedlings were damaged by falling litter. Under the kauri crown, nutrient availability and probably also soil moisture availability are lower than outside the kauri crown (Chapter 2). Our results show that mapau seedlings suffer relatively more than kauri seedlings

56 Growth of seedlings

from this reduced availability of nutrients and water. Nutrient and water availability were much lower in the control plots than in the trenched plots (Tables 4.1 and 4.2), and the reduction of the growth in the control plots, compared to the growth in the trenched plots, was larger for the mapau seedlings than for the kauri seedlings (Figure 4.2). This difference between the two species can probably be explained by a number of qualities of kauri, such as a long leaf life span and a thick cuticle, which enable kauri seedlings to minimise nutrient loss and stand dry conditions well, but which also result in relatively low growth rate (Bieleski 1959; Peterson 1962a; Barton 1982; Bond 1989). In kauri forest, kauri seedlings can achieve considerable ages while waiting in the understory for favourable growing conditions, such as the creation of a canopy gap (Ogden et al. 1987). A study of seedling numbers below and outside kauri indicated that outside kauri the survival of kauri seedlings is lower than the survival of angiosperm seedlings while below kauri their survival is similar (Chapter 3). The results of the present study confirm these findings and suggest that the conditions for kauri seedlings which wait for events like tree fall and the creation of a canopy gap, are relatively more favourable below kauri than outside the kauri crown. Because of the low nutrient availability and the low soil moisture availability under the kauri crown, there the kauri seedlings will not be as easily overgrown by angiosperm seedlings as they will be outside the crown.

57 Chapter 5

58 Tannin addition

Chapter 5

Short-term and long-term effects of tannins on nitrogen mineralisation and litter decomposition in kauri forests

Kauri (Agathis australis (D. Don) Lindl.) occurs naturally in the warm temperate forest of northern New Zealand where it grows mixed with angiosperm tree species. Below mature kauri trees thick organic layers develop in which large amounts of nitrogen are accumulated. This nitrogen seems to be inaccessible to plants. While litter quality can explain the low decomposition rate below kauri, it is not known what causes the accumulation of nitrogen. We hypothesised that kauri tannins reduce nitrogen mineralisation and litter decomposition below kauri. We further hypothesised that high tannin concentrations in the soil would increase the availability of dissolved organic nitrogen relative to the availability of inorganic nitrogen. To test these hypotheses a laboratory incubation was carried out for one year. Purified tannins of kauri and of two other common New Zealand tree species were added to samples of the soil organic layer from under a kauri tree. The results suggest that during the first month of incubation the added tannins reduced nitrogen availability by sequestering proteins or by stimulating nitrogen immobilisation. In the long-term, the reduced nitrogen release which was found following tannin addition seems attributable to the complexation of proteins by tannins. It further appeared that the addition of tannins did not change the ratio of dissolved organic nitrogen to inorganic nitrogen in the long-term. We conclude that the effect of kauri tannins on nitrogen release offers a good explanation for the accumulation of nitrogen below kauri trees.

Based on: Verkaik, E., Jongkind, A. G., Berendse, F. Short-term and long-term effects of tannins on nitrogen mineralisation and litter decomposition in kauri (Agathis australis (D. Don) Lindl.) forests. Plant and Soil in press.

59 Chapter 5

Introduction

The New Zealand kauri tree (Agathis australis (D. Don) Lindl.) can be regarded as a long- lived pioneer. It establishes early in the succession in relatively open ‘pioneer’ conditions, but trees can live for more than 1000 years (Ogden and Stewart 1995). Kauri occurs naturally in the warm temperate forest of northern New Zealand where it grows mixed with angiosperm tree species. Below mature kauri trees very thick organic layers develop, attaining a thickness of 2 metres or more near the trunk of the tree (Silvester and Orchard 1999). The organic layer 2 contains large amounts of nitrogen – up to 0.65 kg N/m – that seem to be inaccessible to plants (Silvester 2000) and it is assumed that the thick organic layer leads to increased weathering of the underlying soil (Swindale 1957; Jongkind and Buurman 2006). The accumulation of organic matter below kauri can at least partly be explained by its litterfall. Kauri leaves have a thick and extremely heavy cuticle (Barton 1982). Kauri leaf litter accounts for only about 35% of kauri litterfall; over 40% is made up of woody material like cones, branches, and bark (Enright 1999; Silvester and Orchard 1999). One way that plants can affect nitrogen mineralisation is by the production of polyphenols. Earlier studies indicate that polyphenols of kauri play a role in the podzolisation process under kauri (Bloomfield 1957). Phenolics are chemically defined as compounds that possess an aromatic ring bearing a hydroxyl substituent, including their functional derivatives (Figure 1.1). Polyphenols have several or many phenolic hydroxyl substituents (Waterman and Mole 1994). The most abundant polyphenols of woody plants are condensed tannins. Tannins – polyphenolic compounds ranging in molecular weight between 500-3000 Daltons – are present in plant cells in vacuoles, cell walls and in the intercellular spaces (Hättenschwiler and Vitousek 2000; Kraus et al. 2003a). They can leach from plant tissue (Hernes et al. 2001) and thus reach the soil. Nitrogen availability and litter decomposition can be reduced in several ways by polyphenols. Polyphenols, and especially those of relatively high molecular weight like tannins, have the ability to sequester proteins in complexes that are resistant to decomposition. Polyphenols may also inhibit microbial activity by being toxic or by interacting with microbial exoenzymes. Further, especially the compounds of low molecular weight, may act as a carbon source for the growth of microbes, which can lead to the immobilisation of nitrogen. Polyphenols may also reduce decomposition by themselves being resistant to decomposition, or by coating other compounds (Field and Lettinga 1992; Hättenschwiler and Vitousek 2000; Kraus et al. 2003a). In this study we tested four hypotheses. The first was that polyphenols of kauri reduce nitrogen mineralisation. The second hypothesis was that polyphenols of kauri reduce litter

60 Tannin addition

decomposition. Since especially large polyphenol molecules like tannins are able to sequester proteins, we focussed on tannins. Northup et al. (1995; 1998) suggest that polyphenols in plant litter can benefit the plants by altering the availability of dissolved organic nitrogen (DON) and inorganic nitrogen. This would be beneficial for those plants that are able to take up the organic nitrogen (for instance, via ectomycorrhizae). We therefore tested a third hypothesis: a high tannin concentration of kauri litter results in greater availability of DON relative to the availability of inorganic nitrogen. In another study, we measured the tannin concentration in the foliage of a number of common plant species of the mixed forest of northern New Zealand, and it appeared that the gymnosperm rimu (Dacrydium cupressinum) has similar tannin concentrations as kauri (Chapter 6). Further, it is known that the decomposition of kauri foliage is not extremely slow compared to the decomposition of foliage of New Zealand species like rewarewa (Knightia excelsa) or silver tree fern (Cyathea dealbata) (Enright and Ogden 1987). However, especially kauri is known for its extreme accumulation of organic material and nitrogen. Our fourth and final hypothesis was that kauri tannins have stronger effects on nitrogen mineralisation and litter decomposition than the tannins of other common New Zealand tree species. To study the effects of tannins, we purified tannins of kauri and of two other common New Zealand tree species and added them to material from the organic layer of the kauri forest. The experiment was run for a year to study long-term effects.

Material and methods

Tannin purification To obtain purified tannins we used the methods of Schimel et al. (1996) and Yu and Dahlgren (2000). We collected samples of about 400 g fresh green leaves from kauri, (Agathis australis (D. Don) Lindl.), the gymnosperm rimu (Dacrydium cupressinum Lambert) and the angiosperm rewarewa (Knightia excelsa R. Br.) at three sites in the Waitakere Ranges in April and May 2003. The leaves were dried, ground and stored in a freezer. The ground material was washed three times with methylene chloride and extracted three times with 80% aqueous acetone. After each washing or extraction the plant material was separated from the solvent using vacuum filtration. The acetone extractable fraction was roto-evaporated to remove the acetone, and the residue was washed with hexane. The water fraction was concentrated on a roto-evaporator and freeze-dried. The resulting powder was redissolved in 50% aqueous methanol and mixed through a slurry of sephadex LH-20 and 50% aqueous methanol. This slurry was applied to a glass funnel and eluted with 50% aqueous methanol until only low

61 Chapter 5

concentrations of polyphenols were detected in the effluent using the Folin-Ciocalteu method (Waterman and Mole 1994). The sephadex was washed with 80% aqueous acetone to remove the tannin and the eluate was then roto-evaporated and freeze-dried to obtain a powder of purified tannin.

Design of the experiment The purified tannins were added to samples from the organic layer from under a kauri tree. In May 2003 we collected a sample of 6 kg of the humus and fermentation layer at one site in the Waitakere Ranges (NZMS 260-Q11: 498687). The material was air-dried, transported to the Netherlands, stored for about 9 months in the dark at room temperature, sieved (2 mm) and then sub-samples of 5 g were put in Petri dishes 92 mm in diameter and 15 mm deep, 3 ml demineralised water was added and the dishes were pre-incubated in the dark at 25˚C. After two weeks, the five treatments were started. All the Petri dishes received 2 ml demineralised water; for the control treatment, this was the only addition. To distinguish between the process of sequestering of proteins by tannins and the immobilisation of nitrogen by microbes, one treatment was the addition of 0.1 g cellulose per dish (20 mg/g organic material). The other treatments consisted of the addition of 0.1 g tannins from kauri, rimu or rewarewa per dish (20 mg/g organic material). After the additions the Petri dishes were sealed with Para film and incubated in the dark at 25°C. Every month the incubated Petri dishes were aerated by lifting off their covers, were brought back to their starting weight by adding demineralised water and were then resealed. Batches of 5 Petri dishes per treatment (but 10 dishes from the control) were removed from incubation for analysis at various intervals: after two days (referred to as t=0), after one month and after one year.

Measurements After collection of the dishes, the fresh weight of the organic material in each Petri dish was measured, the moisture content was measured in a sub-sample, and then the dry weight was calculated. The C and N concentrations of the organic material were measured with an element analyser (Fison instruments EA 1108) in dried (30°C) samples that had been pulverised with a ball mill. The moisture content (drying at 105°C) and ash content (loss on ignition) were determined and the results of all measurements were corrected for moisture and ash content. Sub-samples of the organic material were extracted with 1 M KCl and the pH of the extracts was measured. The mineral nitrogen concentrations were measured with a segmented flow analyser (Skalar, SAN plus system, the Netherlands), and the organic nitrogen concentration was determined using the method of Yu et al. (1994). Pulverised samples of the organic material collected two days after tannin addition (at t=0) were

62 Tannin addition

extracted with 50% aqueous acetone (50 mg of sample in 10 ml solvent, following Yu and Dahlgren 2000). In these extracts polyphenol concentrations were measured colorimetrically by the Folin–Ciocalteu method (Waterman and Mole 1994) and using tannic acid (Sigma– Aldrich) as standard.

Statistical analyses All statistical analyses were conducted with the SPSS statistical package for Windows (12.0.1). ANOVAs with Tukey tests (significance set at 0.05) with treatment as fixed factor were used to test for differences in polyphenol concentration, nitrogen concentration, C:N ratio, N loss and mass loss between the different treatments. Before running each ANOVA the homogeneity of variance was tested; if variances were not equal, a non-parametric test was used. The non-parametric Kruskall–Wallis test was performed to compare more than two treatments, while the Mann Whitney test was used to compare two treatments.

Results

At the start the polyphenol concentrations in the Petri dishes to which rimu and rewarewa tannins had been added were as low as the concentration in the control Petri dishes (Table 5.1). Only in the Petri dishes to which kauri tannins had been added was the concentration higher than the control, but at 15 mg/g it was still lower than the added 20 mg/g. At the start and during the experiment, the pH-KCl was low in all treatments (Table 5.1). During the experiment, in all treatments inorganic nitrogen was mainly present as ammonium and the concentration of nitrate was always less than 0.1% of the ammonium concentration (data not shown).

Table 5.1. pH in the KCl extract, polyphenol concentration (mg/g), C:N ratio, extractable + - inorganic nitrogen (the sum of NH4 -N and NO3 -N, in µg/g dry weight) and dissolved organic nitrogen (µg/g dry weight) at the start of the experiment in the different treatments. Row values followed by different letters differ significantly (P<0.05).

Control Cellulose Kauri Rimu Rewarewa tannins tannins tannins pH-KCl 2.7a 2.7a 2.7a 2.7a 2.7a Polyphenol concentration 7.6a 4.7a 15.1b 6.6a 7.2a C-N ratio 31.4a 32.8b 32.1ab 32.2ab 31.9ab Inorganic nitrogen 489a 460a 481a 473a 459a DON 383d 362ab 370bc 379cd 356a

63 Chapter 5

100 c control 50 b b c c c ab a b cellulose a 0 kauri rimu -50 rewarewa µg/g dry weight -100 Inorganic nitrogen DON

Figure 5.1. The difference (±SE; in µg/g dry weight incubated material) between the initial + - and the one month values for extractable inorganic nitrogen (NH4 -N and NO3 -N) or dissolved organic nitrogen. Different letters indicate significant differences (P < 0.05) between treatments.

After one month’s incubation, the inorganic nitrogen in the organic material of the control had increased, but the inorganic nitrogen in the kauri tannin and rimu tannin treatments had decreased (Figure 5.1). The amount of inorganic nitrogen in the cellulose and rewarewa tannin treatments had not changed (Figure 5.1). After one month, the amount of DON had decreased in all treatments, and this decrease was larger in the kauri and rimu tannin treatments than in the other treatments (Figure 5.1). After one year, the organic material in all treatments had released inorganic nitrogen and DON (Figure 5.2). Whereas the cellulose treatment had released the same amounts of inorganic nitrogen and DON as the control, the tannin treatments all had released smaller amounts of inorganic nitrogen and DON (Figure 5.2). After one month incubation, due the release of inorganic nitrogen in the control, the ratio of DON to inorganic nitrogen was significantly lower in the control than in the other treatments (Figure 5.3). In the long-term, the ratios of DON to inorganic nitrogen were similar in all treatments (Figure 5.3).

1500 b b b b control a a a 1000 a a a cellulose kauri 500 rimu rewarewa µg/g dry weight 0 Inorganic nitrogen DON

Figure 5.2. The difference (±SE; in µg/g dry weight incubated material) between the + - initial and the one year values for extractable inorganic nitrogen (NH4 -N and NO3 -N) or dissolved organic nitrogen. Different letters indicate significant differences (P < 0.05) between treatments.

64 Tannin addition

1 a a a a a a a a a a bc c 0.8 b b a control 0.6 cellulose kauri 0.4 rimu rewarewa 0.2

0 ratio DON to inorganic nitrogen start one month one year

+ - Figure 5.3. The ratio (±SE) of DON to inorganic nitrogen (the sum of NH4 -N and NO3 -N) at the start, after one month and after one year incubation. Different letters in the same incubation period indicate significant differences (P < 0.05) between treatments.

After one month and also after one year of incubation, mass loss in the control was similar to the mass loss in the tannin addition treatments (Figure 5.4). After one month of incubation, the addition of cellulose had brought about greater mass loss compared to the addition of kauri tannins (Figure 5.4), while after one year the mass loss in the cellulose addition treatment was greater than the mass loss in the control and in treatments with kauri and the rewarewa tannin additions (Figure 5.4).

20 b 15 ab control a a a cellulose 10 kauri 5 b rimu ab ab ab a rewarewa 0

-5 one month one year mass loss (%weight at start)

Figure 5.4. Mass loss (±SE, as % of weight at the start) after one month and after one year of incubation. Different letters in the same incubation period indicate significant differences (P < 0.05) between treatments.

65 Chapter 5

Discussion

Nitrogen mineralisation Since kauri tannin addition resulted in a reduction of net nitrogen release compared to the control (Figures 5.1 and 5.2), our first hypothesis, that kauri tannins can reduce nitrogen mineralisation is confirmed. The results indicate that the added kauri tannins were not toxic for micro-organisms, since adding a toxic compound would have inhibited microbial activity and reduced mass loss (Kraus et al. 2003a). The tannin fraction which we purified is dominated by high molecular weight tannins (Fierer et al. 2001). Research suggests that, besides being able to sequester proteins, these tannins might also act as a carbon source, stimulating nitrogen immobilisation (Kraus et al., 2004). In our study we used cellulose for comparison, and the treatment with this compound showed a different response than the addition of kauri tannins. Just as in the study done by Schimel et al. (1996), the cellulose in our study probably acted as a temporary substrate for the microbial population, thereby stimulating nitrogen immobilisation compared to the control one month after incubation (Figure 5.1). Compared to the control the cellulose treatment increased mass loss (Figure 5.4), probably because of the larger microbial population which was present. On the short term the kauri tannin treatment however, showed no effect on mass loss (Figure 5.4), suggesting that in this treatment nitrogen immobilisation was not as important as in the cellulose treatment. Still, the addition of kauri tannins caused a larger decrease of DON availability than the addition of cellulose (Figure 5.1). Also on the long- term both compounds acted differently, with cellulose stimulating mass loss and the tannin addition showing a larger reduction of nitrogen availability (Figures 5.2 and 5.4). It is commonly found that tannins, especially tannins with high molecular weight, have the ability to complex proteins (e.g.. Bradley et al. 2000; Fierer et al. 2001; Kraus et al. 2004). Kauri tannins appear to have this ability too (Jongkind, unpublished data), and therefore we assume that the reduction of DON availability following kauri tannin addition is mainly attributable to the complexation of proteins by kauri tannins. A reduced availability of DON would decrease nitrogen mineralisation, reducing the release of inorganic nitrogen. In addition, the ongoing consumption of inorganic nitrogen by the microbial population would further reduce the availability of inorganic nitrogen. Both processes can explain the reduced availability of inorganic nitrogen following kauri tannin addition (Figure 5.1). One month after incubation the difference in nitrogen release (the sum of inorganic nitrogen and DON) between the control and the kauri tannin addition was only 0.1 mg/g (Figure 5.1), but during the incubation in the following 11 months this increased to 0.64 mg/g

66 Tannin addition

(Figure 5.2). Our results therefore indicate that the complexation of proteins by tannins can require a long time period.

Litter decomposition Our results did not confirm the hypothesis that added tannins reduced litter decomposition. Decomposition was probably carbon limited in our experiment, as indicated by the stimulation of mass loss by cellulose addition (Figure 5.4). The extra mass that the cellulose treatment lost compared to the control exceeded the mass of the added cellulose. After one year, the cellulose treatment had lost 6.3% more mass than the control (Figure 5.4) but the mass of the cellulose treatment at the start (0.1 g cellulose out of 5.1 g dry weight) only comprised 2% of cellulose. Presumably the addition of cellulose fuelled an increase of the microbial population, which then decomposed more soil organic matter than the microbial population of the control. One of the ways tannins can reduce litter decomposition is by complexing proteins and extracellular enzymes of microbes (Kraus et al. 2003a). The kauri tannins we used appear to be able to complex proteins too (see discussion above), and one might expect that therefore they would also be able to complex extracellular enzymes. Both processes would result in a reduction of litter decomposition after tannin addition. However, we found no reduction of mass loss after tannin addition (Figure 5.4). Fierer et al. (2001) reported similar findings. They added balsam poplar (Populus balsamifera) tannins to soil organic material of balsam poplar vegetation and to soil organic material of alder (Alnus tenuifolia) vegetation. In both soil types the high molecular weight tannins complexed proteins. The soil respiration, however, was only reduced in the alder soil material and not in the poplar soil material. Fierer et al. (2001) suggested that the microbes in the poplar soil material, which are naturally exposed to high tannin inputs, are adapted to high tannin concentrations. Therefore they will not be inhibited as strongly as the microbes in the alder soil material, which are not normally exposed to high tannin inputs. The same argument could be used to explain our results: the soil below kauri trees is naturally exposed to high tannin inputs and therefore the microbial population will be adapted to high tannin concentrations.

Dissolved organic nitrogen The results after one month incubation confirm the hypothesis that tannins change the availability of DON relative to inorganic nitrogen, but the results also show that in the long- term the ratio of DON to inorganic nitrogen did not change (Figure 5.3). In a tannin addition experiment which was run for 10 weeks, Bradley et al. (2000) found no shift in the ratio of DON to inorganic nitrogen either. However, they found no effect of tannin addition on the

67 Chapter 5

DON concentration, whereas in our study the DON concentration was even lowered by the addition of tannins (Figure 5.2). In the study done by Bradley et al. (2000) the concentration of DON was very low compared to the concentration of inorganic nitrogen, probably because of the pre-treatment leaching in their experiment. In our experiment the DON concentration was similar to the concentration of inorganic nitrogen (Table 5.1, Figure 5.3) and both the DON and the ammonium release were reduced by tannin addition (Figure 5.2). As a result, the ratio of organic to inorganic nitrogen was not changed in the long-term.

Kauri tannins versus tannins of other species The hypothesis that kauri tannins show a stronger effect on nitrogen mineralisation than tannins of other species was not confirmed, since the effect of kauri tannins on the availability of inorganic nitrogen was similar to the effects of rewarewa tannins or rimu tannins (Figure 5.1). However, after one month incubation the effect on DON availability was larger in the kauri tannin treatment than in the rewarewa tannin treatment. Other studies have also found that purified tannins of different species act differently (Kraus et al. 2003b), and differences in tannin structure appear important. Since the short-term effects of rewarewa tannin addition, on nitrogen availability and mass loss, were similar to the effects of cellulose addition (Figures 5.1 and 5.4), we are not able indicate if these short-term effects of rewarewa tannins were due to complexation or to immobilisation. It appears that the rimu tannins not only reduced nitrogen availability by immobilisation but also by the complexation of proteins in the short-term, since the rimu tannin treatment showed a larger decrease of nitrogen availability than the cellulose treatment but showed a similar mass loss. In the long term, the effects on nitrogen release and mass loss were similar for the three tannin additions and differed from the cellulose treatment (Figures 5.2 and 5.4), suggesting that in the long-term the complexation of proteins by tannins is the main process affecting nitrogen availability. Two days after tannin addition, the polyphenol concentrations for all three kinds of tannins were low compared to the added quantities of 20 mg/g (Table 5.1). Similar results have been found in other studies. Schofield et al. (1998) added purified sorghum tannins to soil samples and used a variety of extraction solutions, but were unable to measure any of the added tannins. Two weeks after the addition of tannins to humus Bradley et al. (2000) did not detect any tannin in the humus leachates. These studies and our findings show that the added tannins quickly become unavailable for extraction, indicating that the tannins bind to the substrate. Although these tannins are unavailable for extraction, the results show that they can still influence nitrogen mineralisation.

68 Tannin addition

Nitrogen accumulation in kauri forest In our study the effects of kauri tannins on nitrogen availability were assessed under laboratory conditions. In kauri forest the effects of tannins might be different, due to for instance the presence of soil mesofauna or to fluctuating environmental conditions (Kraus et al. 2003a). However, our study shows that potentially kauri tannins have a large effect on nitrogen availability. One year after the start of the experiment the difference in nitrogen release (the sum of inorganic nitrogen and DON) between the control and the kauri tannin addition was 0.6 mg nitrogen (Figure 5.2). This suggests that 20 mg of kauri tannins can -2 sequester 0.6 mg nitrogen. In kauri forest the annual leaf litterfall is 54-277 g m (Enright 1999), which, given the tannin concentrations measured in another study (Chapter 6) gives an -2 -1 input of about 6-33 g kauri tannins m y . If these kauri tannins complexed nitrogen to the same extents as they did in our experiment, this would result in a nitrogen sequestration of -2 -1 0.18-0.99 g nitrogen m y . During the 600-700 years’ lifetime of a kauri tree (Ahmed and -2 Ogden 1987), a sequestration of 0.09-0.5 kg nitrogen m would be possible which is comparable to the amount of nitrogen in the organic layer of a mature kauri forest (Silvester 2000). The purified tannins of the other two species had similar long-term effects on nitrogen release as did kauri tannins (Figure 5.2). This emphasises that also other tree species in New Zealand kauri forest have the ability to sequester proteins. We propose that the exceptional accumulation of nitrogen under kauri can be explained by the combination of the high attainable age of kauri and by its ability to occupy certain positions in the landscape for several generations of trees – a possible time span of thousands of years (Ogden and Stewart 1995).

69 Chapter 6

70 Variation in tannin concentration

Chapter 6

Variation in tannin concentration among species of New Zealand kauri forest and the effects of kauri tannins on litter decomposition and nitrogen mineralisation

Plants might enhance their own fitness by improving soil quality or by making the soil less favourable for competing species. In the latter strategy, tannins in plant foliage might be important. In a study of the New Zealand kauri tree (Agathis australis), whose massive litter accumulation has a huge impact upon the soil, we tested the hypotheses that: i) the foliar tannin concentrations are higher in kauri than in other species of the kauri forest, ii) nitrogen mineralisation rate and litter decomposition rate are lower in kauri leaves with a high tannin concentration than in leaves with a lower tannin concentration, and iii) variation in the tannin concentration of kauri foliage is heritable. To compare the tannin concentrations in plant species of the kauri forest we collected foliage from ten plant species. To test if tannin concentration in kauri leaf litter affects nitrogen mineralisation and litter decomposition we conducted a litterbag experiment with two kinds of kauri litter differing in tannin concentration. To test if genetic variation of the tannin concentration in kauri foliage exists we collected seeds of kauri individuals in three forest areas, raised seedlings from them and compared their tannin concentrations. Our results confirmed that the tannin concentration in kauri foliage was higher compared to most other common species of the kauri forest, and also confirmed that there is genetic variation in tannin concentration. Further, it was found that nitrogen is released slowly from kauri leaf litter, what might have consequences for the survival of seedlings below kauri. Nitrogen loss was similar but the increase in nitrogen concentration was different for the two kinds of kauri litter which differed in tannin concentration. Therefore, the litterbag experiment provided no conclusive evidence in favour or against an effect of kauri tannins on nitrogen mineralisation.

Based on: Verkaik, E., Berendse, F. Variation in tannin concentration among species of New Zealand kauri forest and the effects of kauri tannins on litter decomposition and nitrogen mineralisation. Submitted.

71 Chapter 6

Introduction

The chemical and physical properties of soils are influenced by plants and the changes that these plants bring about in the soil might alter the competition among plant species (e.g. Van Breemen 1993; Berendse 1994a). If plant species affect soil properties and the changes in the soil properties have positive effects on the fitness of that plant species, a positive feedback can develop between the species and the induced changes in the soil (Berendse 1998; Van Breemen and Finzi 1998). The impact of the New Zealand kauri tree (Agathis australis) on the soil might be an example of such a positive feedback. Under the New Zealand kauri tree very thick organic layers develop, possibly becoming over 2 meters thick near the stem of mature trees (Silvester and Orchard 1999). This organic layer contains large amounts of 2 nitrogen, up to 0.65 kg N/m , which seem to be inaccessible to plants (Silvester 2000). The thick organic layer leads to increased weathering of the underlying soil (Swindale 1957; Jongkind and Buurman 2006) and soil fertility appears to be lower in the organic layer below the crown of a kauri tree than in the mineral soil outside the crown (Chapter 2). Kauri is well adapted to grow on nutrient poor soils, since it has a low nutrient requirement due to the long life span of its leaves (Peterson 1962a; Silvester 2000). Seedlings of kauri are able to survive under mature kauri trees (Ogden and Stewart 1995). An earlier study suggests that the low soil fertility below the kauri crown hampers angiosperm seedlings relatively more in their growth than kauri seedlings (Chapter 3). Polyphenols and especially larger polyphenol molecules like tannins are known to be able to complex nitrogen. High polyphenol concentrations in plant litter can also reduce the rate of litter decomposition and so usually result in the accumulation of a thick layer of humus and litter (mor humus) and low rates of nitrogen mineralisation in the soil (Hättenschwiler and Vitousek 2000; Kraus et al. 2003a). We purified kauri tannins and in a laboratory experiment added these tannins to organic material. It appeared that kauri tannins have the ability to slow down nitrogen mineralisation (Chapter 5). In the present study we tested four hypotheses: 1) tannin concentrations in kauri foliage are high compared to concentrations found in other common species of the kauri forest; 2) decomposing kauri leaves with high tannin concentrations release less inorganic nitrogen than kauri leaves with lower tannin concentrations; 3) kauri leaves with high tannin concentrations decompose more slowly than kauri leaves with lower tannin concentrations. To assess whether soil-plant feedbacks are driven by natural selection, it is not only necessary to test if plants affect their own fitness via soil processes, but it is further necessary to establish that individuals within a species vary genetically in the attributes that influence soil properties (Van Breemen and Finzi 1998). The last hypothesis we tested (4) is that the

72 Variation in tannin concentration

variation in the tannin concentration of kauri foliage is heritable and kauri trees are thus able to confer the trait ‘high tannin concentration’ to their offspring. To test for differences in tannin concentration between plant species (hypothesis 1) we collected foliage of kauri trees and of nine other common plant species of the kauri forest and measured tannin concentrations. To test for effects of tannin concentration on nitrogen mineralisation and litter decomposition (hypotheses 2, 3) we incubated two kinds of kauri leaf litter, differing in tannin concentration, in litterbags in kauri forest. Genotypic variation (hypothesis 4) was analysed by collecting cones of kauri individuals in three forest areas, raising seedlings from the seeds in the cones and comparing the tannin concentrations among the seedlings.

Methods

Comparison of tannin concentration among species To test for differences between the polyphenol and tannin concentrations of leaves of kauri trees and leaves of other common plant species of the kauri forest, we sampled plants at five different sites within the Waitakere Ranges, New Zealand (Topomap 260,Q 11 & Pt. R11/ 497689, 507701, 452781, 464784, 462777). Leaves of the following ten species were collected: kauri; rimu (Dacrydium cupressinum), a large emergent coniferous tree that is common in kauri forest; rewarewa (Knightia excelsa), an emergent broadleaved tree with litter that decomposes slowly (Enright and Ogden 1987); silver tree fern (Cyathea dealbata), a common tree fern with slowly decomposing leaf litter; a palm (Rhopalostylis sapida); three common broadleaved understory tree species (, Myrsine australis and Coprosma arborea); and two species that often dominate the herb layer on the kauri litter layer (kauri grass, Astelia trinervia and kiekie, Freycinetia banksii). The leaves were collected in April 2003. Per site all species were collected on the same day. At each site we collected leaf material from 2-5 mature plants per species. Rhopalostylis sapida has large pinnate leaves, so instead of collecting total leaves we collected individual leaflets; for Cyathea dealbata the primary pinnae were collected. For most species, leaves were picked fresh from plants, when necessary using a pruning pole. At four of the sites the crowns of kauri and Dacrydium cupressinum trees were often beyond reach, so instead we picked green leaves of kauri and Dacrydium cupressinum from the forest floor or from twigs that had fallen on the forest floor. The material collected from the individual plants was bulked per site and species to give one composite sample of about 30 g fresh leaf material. The collected leaf material

73 Chapter 6

was air and oven dried (max. temp. 35 °C) and stored in a cooler (4 °C) until transport to The Netherlands, where it was frozen until chemical analysis. A small test was conducted to ascertain whether there was any difference in the chemical composition of green kauri leaves picked from trees and green kauri leaves collected from the forest floor. At three sites, we collected twigs with green leaves by picking them from kauri trees using a pruning pole. We distributed the collected twigs over three samples of five twigs. The leaves of the first sample were dried immediately after collection. The second sample was left for 5 days on the ground under a vegetation of tea tree (Leptospermum scoparium) bush, after which the leaves were picked from the twigs and dried, and the third sample was left on the forest floor until the leaves on the twigs were beginning to turn yellow (14 days) after which the twigs were collected and the leaves dried and analysed using the methods described below. There were no significant differences (P < 0.05) in the nitrogen, phosphorus, polyphenol or tannin concentrations between the fresh leaves and those collected from the forest floor 14 days later. Therefore we assumed that the nutrient and tannin concentrations of green kauri leaves collected from the forest floor would give a good indication of the concentrations in the green leaves on the trees.

Litterbag experiment To test the hypotheses that kauri leaves which differ in tannin concentration have different nitrogen mineralisation and decomposition rates, we collected kauri leaves in two forest areas. In June 2002, leaves of kauri which looked if they had only recently been abscised, with a gold-brown colour, were collected from the forest floor at five different sites within the Waitakere ranges and at three different sites within Thomson Kauri Grove Scenic Reserve (Kaipara flats, topographic map 260 Q09/491305). The field-moist material was incubated in mesh bags, of 10 x 10cm with a mesh size of 1 x 1mm. The mesh bags were incubated below and outside (5 to 15 m away from) the crown of 5 different mature kauri trees, and per bag 7 (±0.05) g material was incubated. The bags were pinned to the forest floor and were covered with a thin layer (0.5-3 cm) of leaf litter. For each kind of litter 20 mesh bags were incubated under the crowns of the 5 kauri trees (4 bags per tree) and 20 mesh bags outside the crown of the 5 trees (4 bags per tree). For each location and each litter species half of the bags was collected from the field after 12 months incubation and the remaining bags were collected after 28 months. At the start of the experiment for each litter species 20 reference samples of 7 (±0.05) g leaf litter were taken. The reference samples were immediately oven dried (24 hours at 35°C) and weighted. The samples which had been incubated in the forest were first cleaned, by removing with a pair of tweezers soil particles from the remaining litter, and were then oven dried and weighted. Then the samples were transported to The Netherlands where

74 Variation in tannin concentration

they were stored until chemical analysis. The reference samples were stored for three months at room temperature while the samples which were collected after 12 and after 28 months were stored in a freezer for 5 and 6 months, respectively.

Variation in tannin concentration between kauri seedlings In March 2002 we collected kauri seeds and kauri seed cone material in three forest areas to investigate what part of the variation in the tannin concentration of kauri foliage is heritable. In the Waitakere Ranges, seed cones were collected from five kauri trees. We collected one cone (two trees), two cones (two trees) and four cones (one tree) from each tree. In Thomson kauri reserve we collected one cone per tree from the forest floor under four different trees. In forest on the Awhitu peninsula (255 Awhitu central road, topographic map 260-Q12 & R12/529550) we collected seeds from the forest floor under four different kauri trees. The kauri seeds were air dried for several days and flown to The Netherlands where they were sown in pots with a mixture (50% of the volume) of upgraded black (Naturado) and perlite, and kept in a greenhouse. The seedlings were potted on twice (at 9 and 15 months) in larger pots containing the same mixture of peat and perlite. After 20 months the seedlings were harvested, and for each seedling the length of the shoot and the dry weights of the shoot and root were measured and the leaves were counted.

Chemical analyses To analyse the elemental composition of the foliage of the kauri trees, of the foliage of the nine other species of the kauri forest and of the samples of the litterbag experiment, the material was ground and digested with sulphuric acid, salicylic acid, hydrogen peroxide and selenium (Temminghoff et al. 2000). The nitrogen and phosphorus concentrations in the digest were determined spectrophotometrically (Skalar, SAN plus system or AAS). To analyse total C, tannin and polyphenol concentrations in all plant material and to analyse total N in the foliage of the seedlings, the material was pulverised with a ball mill and total C and N concentrations were measured with an element analyser (Fison instrument EA 1108). Moisture concentration (by drying at 105 ºC) and the ash content (by loss on ignition) were determined and the results of all chemical analyses were corrected for moisture and ash concentration. The concentrations of polyphenols and tannins were measured colorimetrically in 50% aqueous acetone extracts (Yu and Dahlgren 2000) using the Folin-Ciocalteu and the acid butanol methods respectively (Waterman and Mole 1994). The tannin concentrations of the foliage of kauri, Dacrydium cupressinum and Knightia excelsa were measured using their purified tannins as standards. For the other seven plant species, the tannins of kauri, Dacrydium cupressinum and Knightia excelsa were used as standard. Since the calculated

75 Chapter 6

tannin concentrations did not differ much between these three standards, here we only report the calculated tannin concentrations using kauri tannins as standard. For all species, tannic acid (Sigma-Aldrich) was used as the standard when calculating polyphenol concentrations. For tannin purification we used the methods of Yu and Dahlgren (2000) and Schimel et al. (1996), as described in Chapter 5.

Statistical analyses All statistical analyses were conducted with the SPSS statistical package for Windows (12.0.1). To analyse the data of the litterbag experiment ANOVA were performed whereby ‘location’ (under or outside the kauri crown) and ‘litter type’ were considered fixed factors and ‘kauri tree’ was considered a random factor. Interactions between the factors were not included. ANOVA followed by Tukey tests were performed to test for differences in carbon, nitrogen, phosphorus, tannin and polyphenol concentration in foliage between the ten species. An ANOVA with a nested design was used to analyse the seedling data. To test for differences between forest areas, the random factors ‘cone’ and ‘tree’ were considered nested within each ‘area’; to test for differences between trees within an area, ‘cone’ was nested within ‘tree’. Before running each ANOVA, homogeneity of variance was tested; if variances were not equal, non-parametric tests were used. The non-parametric Kruskal-Wallis test was used to compare several unrelated groups, the non-parametric Mann-Whitney test was used for two unrelated groups and the non-parametric Wilcoxon test was used for the paired samples of the litterbag experiment. The broad heritability of the tannin concentration, the genotypic variation as proportion of the phenotypic variation (Griffiths et al. 1996), was calculated using the variance components analysis in SPSS. For each forest area the variation in tannin concentration among seedlings of different trees (the added variance) was expressed as proportion of the sum of the variation among trees (added variance) and within trees (error variance). This number was then multiplied by two, assuming that seedlings from the same tree had half their genes in common. Pearson’s correlation coefficients were calculated to test whether the nitrogen concentration of the seedlings correlated with biomass and tannin concentration.

Results

Comparison of tannin concentration between species The tannin and polyphenol concentrations were much higher in leaves of kauri than in the leaves of all but one of the nine other species collected in the Waitakere Ranges (Table 6.1).

76 Variation in tannin concentration

They were higher in kauri than in species with mesophyllous leaves (Melicytus ramiflorus and Myrsine australis) and species with hard sclerophyllous leaves (Rhopalostylis sapida, Cyathea dealbata and Knightia excelsa). The only species to have higher tannin concentration than kauri was the single other coniferous species sampled: Dacrydium cupressinum.

Table 6.1. Mean (±SE) carbon, nitrogen, phosphorus, polyphenol and tannin concentrations (mg/g dry weight) in leaf material of ten plant species. Column values followed by different letters are significantly different (P<0.05). Species Carbon Nitrogen Phosphorus Polyphenol Tannin Kauri, Agathis australis 565 ± 4de 6.21 ± 0.41ab 0.44 ± 0.02bc 172 ± 10e 119 ± 16d Dacrydium cupressinum 578 ± 6e 7.59 ± 0.38bc 0.52 ± 0.04cd 191 ± 8e 180 ± 12e Knightia excelsa 535 ± 3bc 4.99 ± 0.36a 0.27 ± 0.02a 93 ± 5cd 40 ± 4c Cyathea dealbata 492 ± 14a 12.11 ± 0.71d 0.59 ± 0.05cd 119 ± 16d 24 ± 5c Rhopalostylis sapida 505 ± 6a 13.68 ± 0.48d 0.86 ± 0.03e 25 ± 1ab 5 ± 1a Coprosma arborea 503 ± 8a 12.09 ± 0.71d 0.75 ± 0.05de 81 ± 4c 28 ± 1c Melicytus ramiflorus 512 ± 6ab 19.85 ± 1.10e 1.04 ± 0.07e 24 ± 2ab 3 ± 0a Myrsine australis 548 ± 7cd 9.33 ± 0.46c 0.53 ± 0.02c 81 ± 6c 34 ± 4c Astelia trinervia 514 ± 10ab 4.67 ± 0.33a 0.30 ± 0.02a 33 ± 6b 10 ± 3b Freycinetia banksii 513 ± 5ab 9.30 ± 0.36c 0.61 ± 0.06d 14 ± 4a 4 ± 1a

Litterbag experiment The nitrogen concentrations of the senesced kauri leaves used in the litterbag experiment (Table 6.2) were much lower than the nitrogen concentrations of the fresh kauri leaves (Table 6.1). The senesced leaves which we collected in the Waitakere Ranges had higher tannin and polyphenol concentrations and lower nitrogen concentrations than the leaves from Thomson kauri reserve (Table 6.2).

Table 6.2. The initial tannin, polyphenol and nitrogen concentrations (±SE, in mg/g dry weight); the nitrogen loss after 28 months incubation (±SE, in mg/g incubated litter); and the C:N ratios (±SE) at the three time steps, in kauri leaf litter from the Waitakere Ranges and from Thomson kauri reserve. Row values followed by different letters are significantly different (P < 0.05). Waitakere Thomson Tannin concentration 66.8 ± 1.2b 45.5 ±1.1a Polyphenol concentration 106.9 ± 1.4b 94.6 ±1.2a Nitrogen concentration 3.92 ± 0.06a 4.90 ± 0.08b Nitrogen loss after 28 months 1.04 ± 0.25 1.53 ± 0.33 C:N ratio start 137 ± 2b 108 ± 2a C:N ratio after 1 yr 109 ± 4b 91 ± 1a C:N ratio after 28 months 74 ± 1b 66 ± 2a

77 Chapter 6

80

60 12 months Under 12 months Outside 40 28 months Under 28 months Outside 20 loss as % of amount at start at amount of as % loss 0 Waitakere Thomson Waitakere Thomson

Mass loss N-loss

Figure 6.1. Mass-loss and nitrogen-loss (±SE, as percentages of amounts at the start) for kauri leaves from the Waitakere Ranges and Thomson kauri reserve incubated under and outside the crown of kauri trees.

During the first year of incubation, the Waitakere leaves and the Thomson leaves both lost about 35% of their mass and lost about 11% of their nitrogen content (Figure 6.1). After 28 months of incubation, mass loss had increased to about 63% for both kinds of kauri leaves. Nitrogen loss had increased to 31% for the Thomson leaves and this loss was about 5% lower for the Waitakere leaves (Figure 6.1), but not significantly so (Table 6.3). Due to the higher relative mass loss than nitrogen loss, the nitrogen concentrations increased in both kinds of litter and their C:N ratio’s decreased (Table 6.2, Figure 6.2). The increase in nitrogen concentration was higher in leaves from the Waitakere Ranges than in leaves from Thomson, 28 months after incubation (Table 6.3, Figure 6.2). Nitrogen loss and mass loss were similar for leaves incubated below and leaves incubated outside the crown of kauri trees (Table 6.3, Figure 6.1).

Variation in tannin concentration between kauri seedlings We found no significant differences in tannin concentration between kauri seedlings grown from different cones of the same tree (data not shown). Differences in tannin concentration between seedlings originating from different trees but from the same forest area were not significant for two forest areas but were significant for the seedlings originating from Thomson reserve (data not shown). For the seedlings from Thomson reserve we found a broad heritability (Griffiths et al. 1996) of 74% for the tannin concentration while for the other two areas we found no significant broad heritability. The seedlings from Thomson

78 Variation in tannin concentration

reserve had significantly higher tannin concentrations than the Waitakere seedlings, even when two Thomson seedlings with an exceptionally high tannin concentration (average value of 141 mg/g) were excluded from the analyses (Table 6.4).

Table 6.3. Litterbag experiment. Results of the ANOVA for the factors ‘location’ (below or outside the kauri crown) and ‘leaf origin’ (kauri leaves from the Waitakere Ranges or from Thomson kauri reserve). The P-values are shown and values indicated in bold indicate a significant (P < 0.05) effect of the factor on the parameter.

12 Months 28 Months Factors→ Location Leaf origin Location Leaf origin Parameters↓ Mass-loss 0.711 0.389 0.173 0.951 N-loss 0.383 0.347 0.303 0.544 Increase N- 0.265 0.190 0.514 0.006 concentration

The nitrogen concentration and biomass differed between seedlings with different provenances (Table 6.4). On average, bigger seedlings had lower nitrogen concentrations than smaller seedlings and there was a significant negative correlation between foliar nitrogen concentration and the seedling biomass (r = -0.62, P < 0.01). The differences in nitrogen concentration and biomass between seedlings with different provenances could not account for the difference in tannin concentration which we found between the areas, since tannin concentration was not significantly correlated with seedling weight, nitrogen concentration or C:N ratio.

2.5 leaves Waitakere * 2 leaves Thomson 1.5 1 0.5 0 12 months 28 months Increase in N-concentration Increase

Figure 6.2. Litterbag experiment. Increase of nitrogen concentration (±SE, calculated as nitrogen concentration divided by the nitrogen concentration at the start) for kauri leaves from the Waitakere Ranges and from Thomson kauri reserve. * indicates a significant difference (P<0.05) between the two kinds of kauri leaves.

79 Chapter 6

Table 6.4. The number of seedlings analysed per forest area and the mean (±SE) biomass (g/plant), foliar nitrogen concentration (mg/g dry weight) and tannin concentration (mg/g dry weight) in those seedlings. Column values followed by different letters are significantly different (P<0.05). Forest area Seedlings Biomass Nitrogen Tannin analysed Waitakere 63 4.5 ± 0.2 a 7.6 ± 0.2 b 101 ± 1 a Thomson 15 6.0 ± 0.4 b† 7.0 ± 0.5 b† 112 ± 4 b† Awhitu 13 5.6 ± 0.3 b 5.8 ± 0.4 a 102 ± 4 ab † two seedlings with a high tannin concentration, originating from one tree, were excluded from the analysis.

Discussion

Variation in polyphenol and tannin concentrations among plant species of the kauri forest When we tested the hypothesis that kauri has high concentrations of polyphenols and tannins compared to other species we found that the polyphenol and tannin concentrations in kauri leaves were indeed higher than in many common plant species within the kauri forest (Table 6.1), although not as high as in Dacrydium cupressinum, an other coniferous species. Given that the foliar tannin and polyphenol concentrations of woody species can range up to 25% and 40% dry weight respectively (Kuiters 1990; Kraus et al. 2003a), the tannin and polyphenol concentrations we found in kauri foliage, of 12% and 17% dry weight, were not exceptionally high.

Differences in mineralisation and decomposition between kauri leaves differing in tannin concentration Nitrogen concentrations of the senesced kauri leaves used in the litterbag experiment were similar to the concentrations measured in other studies for kauri leaf litter (Enright and Ogden 1987; Silvester 2000; Enright 2001). The nitrogen concentration of kauri leaf litter is low compared to the nitrogen concentration reported for leaf litter of warm temperate evergreen forests or subtropical evergreen forests world wide (Vogt et al. 1986; Enright 2001). We only observed nitrogen loss and no net nitrogen immobilisation during the experiment (Figure 6.1). Similar results were found for other slowly degradable litter types (Bosatta and Staaf 1982) and the low availability of easily degradable carbon sources might explain this finding (Kooijman and Besse 2002). The low content of easily decomposable carbon probably limited

80 Variation in tannin concentration

microbial growth already early in the decomposition process, resulting in an absence of net nitrogen immobilisation (Kooijman and Besse 2002) The mass loss being higher than the loss of nitrogen, and the relatively high C:N ratio of kauri leaf litter suggest that nitrogen will accumulate below kauri (Enright and Ogden 1987; Vitousek et al. 2002). The results of an earlier laboratory experiment showed that kauri tannins can affect nitrogen release (Chapter 5). In the present study we used two kinds of litter differing 50% in tannin concentration (Table 6.2) to study the effects of kauri tannins on nitrogen release in the field. While we found no significant difference in nitrogen loss, we did find a stronger increase in nitrogen concentration in the litter of the Waitakere Ranges than in the litter of Thomson (Table 6.3, Figure 6.2). This difference might be explained by the higher tannin concentration in the litter of the Waitakere Ranges. However, also the initial nitrogen concentrations differed between the two kinds of litter (Table 6.2), and an alternative explanation for the stronger increase of the nitrogen concentration in the litter of the Waitakere Ranges can be its lower nitrogen concentration at the start of the experiment (Bosatta and Staaf 1982). Therefore, the present litterbag experiment gives no conclusive evidence for an effect of kauri tannins on nitrogen mineralisation in the field. More insight in the fate of nitrogen during decomposition in the field might be provided by a litterbag 15 experiment using N-labelled litter (Zeller et al. 2000), while molecular techniques and nuclear magnetic resonance (NMR) can give information on the fate of tannins (Hernes et al. 2001; Lorenz and Preston 2002). Techniques capable of detecting and quantifying protein- tannin complexes are still in need (Kraus et al. 2003a). The average mass loss of 35% which we measured after one year incubation (Figure 6.1) was close to the 29% mass-loss which Enright and Ogden (1987) measured in a litterbag experiment with kauri leaf litter. Like in our laboratory incubation experiment (Chapter 5), also in the present experiment we found no effect of tannins on the rate of mass loss (Table 6.3, Figure 6.1). Apparently the microbial population in kauri forest is very well adapted to the presence of tannins. May be therefore, following the argumentation of Fierer et al. (2001) we found no effect of tannins on the decomposition rate. The results of the litterbag experiment further suggest that the accumulation of organic material is not caused by the dry and nutrient poor conditions below kauri, since mass loss rates were similar for litterbags incubated below and outside the crown of kauri trees (Table 6.3, Figure 6.1). An explanation for the extreme accumulation of organic material under kauri might be the woody nature of its litterfall and the significant amount of dead bark that falls close to the stem (Silvester and Orchard 1999).

81 Chapter 6

Variation in tannin concentration between kauri seedlings In order to test the hypothesis that the variation in tannin concentration in kauri foliage is heritable, we raised a number of kauri seedlings under equal conditions. The differences in the foliar tannin concentration of the seedlings from the three provenances (Table 6.4) and between different trees within Thomson reserve which we found can therefore –at least partly– be attributed to genetic differences. For the seedlings of Thomsom reserve a broad heritability of 74% was found for the tannin concentration in the foliage. The heritability of a trait changes from population to population and from environment to environment (Griffiths et al. 1996). A comparison of the nutrient concentrations in the leaves of our seedlings with nutrient concentrations of seedlings collected in kauri forests (Peterson 1962b) shows that nutrient availability in our experiment was similar to the nutrient availability in kauri forests. The seedlings which we used were grown from seeds collected in kauri forests and stem from only a small number of kauri trees. We therefore expect that, like in our experiment, also in kauri forests genotypic variation in tannin concentration between seedlings will be present. Genotypic variation in tannin concentration has also been found in other plant species (see for instance Kraus et al. 2003a and Hättenschwiler et al. 2003 and references therein). An alternative explanation for the variation in tannin concentration which we found is a maternal effect, as a result of which the seedling’s tannin concentration is affected not only by its genotype but also, via the seed, by the phenotype of the mother. It is known that seed weight does affect the height of kauri seedlings (Barton 1982). In our study seed weight was not measured, but we did find differences in biomass (Table 6.4). However, we found no correlation between seedling tannin concentration and seedling biomass (r = -0.087, P = 0.41) and this makes an effect of seed weight on tannin concentration of the seedling less likely.

Consequences for kauri ecology The slow release of nitrogen which we measured in our litterbag experiment indicates that one explanation for the accumulation of nitrogen below kauri trees is the nature of its leaf litter (Enright and Ogden 1987; Silvester 2000; Enright 2001). The slow release of nitrogen from kauri litter might affect the fitness of kauri, since an earlier study indicated that the low soil fertility below kauri hampers angiosperm seedlings relatively more in their growth than kauri seedlings (Chapter 3). While a laboratory incubation study strongly suggested that tannins contribute to the slow release of nitrogen from kauri litter (Chapter 5), the results of our litterbag experiment provide no conclusive evidence in favour or against that hypothesis.

82 Variation in tannin concentration

Our measurements of tannin concentration in plant foliage indicate that kauri has not an extremely high tannin concentration. Therefore we suggest that, if kauri tannins have the ability to sequester proteins, it is the combination of kauri’s tannin content, with kauri’s ability to occupy certain positions in the landscape for generations of trees (Ogden and Stewart 1995), that can explain the huge accumulation of nitrogen below kauri. The results of the seedling experiment suggest that the variation in tannin concentration is partly heritable and therefore, if there is a positive effect of kauri tannins on the fitness of kauri trees, natural selection for kauri trees with high tannin concentrations is possible.

83 Chapter 7

84 General discussion

Chapter 7

General discussion

In this thesis I focused on the hypothesis that kauri tannins can reduce nitrogen mineralisation in the soil below the crown of a kauri tree and thus create an environment where kauri seedlings are better able to compete with angiosperm seedling, resulting in natural selection for kauri trees with high tannin concentration. In the foregoing chapters, I presented a number of studies which were conducted to test this hypothesis. Here I will integrate the findings. First I will discuss the influence of kauri trees on the environmental conditions in kauri forest, and then I will describe the effects on seedlings. In the last paragraph I present the conclusions.

85 Chapter 7

The environmental conditions in kauri forests

It is generally accepted that the growth of kauri trees in kauri forests is limited by the availability of nitrogen (e.g. Silvester 1978; Ecroyd 1982; Enright and Ogden 1987; Silvester 2000; Enright 2001). However, some studies indicate that their growth is normally limited by phosphorus availability as well. Peterson studied the growth of kauri seedlings on nutrient solutions and studied their growth in ten kauri forests (Peterson 1961; Peterson 1962a; Peterson 1962b). In addition, he compared the nutrient concentrations of mature kauri trees growing in fertilized stands with those in natural forests. Both studies suggest a widespread deficiency of nitrogen and phosphorus in many kauri forests (Peterson 1962b). A fertilizer trial in another kauri forest led to the same conclusion. In this experiment (Madgwick et al. 1982; Barton and Madgwick 1987) the addition of nitrogen increased growth but the addition of a complete fertilizer, including nitrogen and phosphorus, resulted in an additional increase. I collected kauri leaves from mature trees in three forest areas and found N-P ratios of 14-18 (Table 7.1), also indicating a co-limitation of growth by nitrogen and phosphorus (Koerselman and Meuleman 1996; Güsewell 2004).

Table 7.1. Mean N-P ratios (±SE) and nutrient concentration (% of dry weight ±SE) in leaves of mature kauri trees, measured according to the methods described in Chapter 6. The kauri leaves were collected at five different sites in the Waitakere Ranges, and at three different sites in Thomson reserve and in the forest at the Awhitu Peninsula. N-P ratio N P K Ca Mg Na Waitakere 14.3±0.9 0.62±0.04 0.044±0.002 0.72±0.05 0.64±0.07 0.14±0.01 0.25±0.05 Thomsons 15.7±0.8 0.94±0.03 0.060±0.004 0.73±0.09 0.76±0.12 0.12±0.02 0.25±0.03 Awhitu 17.6±1.8 0.79±0.08 0.047±0.010 0.75±0.10 0.66±0.06 0.18±0.01 0.31±0.11

Kauri forests can be regarded as a mosaic of vegetation patches. Patches of mature kauri trees and patches dominated by other tree species (Chapter 1; Bieleski 1979; Ogden and Stewart 1995; Burns and Leathwick 1996). My study of the nutrient concentrations in leaves of the shrub hangehange (Geniostoma rupestre var. ligustrifolium) shows that nutrient availability is not distributed evenly within this mosaic. Nitrogen, phosphorus and calcium availability appeared to be lower below than just outside the kauri crown (Chapter 2, Figure 2.1). This confirms previous suggestions that the soil below the kauri tree has a low fertility (Swindale 1957; Beveridge 1975; Yeates et al. 1981; Ecroyd 1982; Silvester 2000). One explanation for the deficiency of phosphorus in kauri forests is the fact that most kauri forests occur on deeply weathered soils (Molloy 1993). While newly deposited primary substrates, like volcanic material or glacial debris, usually contain a high quantity of available

86 General discussion

phosphorus, this quantity decreases over time due to binding in physically or chemically protected forms, and due to leaching of small amounts during weathering (Burnham 1989; Vitousek and Farrington 1997; Güsewell 2004). The thick organic layers beneath kauri trees probably contribute to the weathering of the soil (Bloomfield 1953; Swindale 1957; Jongkind and Buurman 2006). Furthermore, the first European settlers in New Zealand logged much of the kauri forest but spared kauri forests situated on poor soils and hilly terrain (Reed 1954; Mc Kelvey and Nicholls 1959; Beever 1981; Halkett and Sale 1986). This is why the remaining kauri forests are situated on nutrient poor and deeply weathered soils. Part of the phosphorus which comes available by weathering is bound in organic form (Vitousek and Farrington 1997). The strong negative correlation (Table 2.3) which I found between the thickness of the organic layer below kauri and the nutrient concentration in leaves of hangehange (Geniostoma rupestre var. ligustrifolium) suggests that phosphorus accumulates in the organic layer below kauri. The correlation further suggests that this phosphorus is unavailable for plants (Chapter 2). Other studies also show that phosphorus accumulates in the organic layer below kauri. In a site of kauri forest the annual return of -2 phosphorus in litterfall was estimated to be 0.31 g m (Enright 2001), which is similar to values reported for warm temperate and for subtropical forests worldwide (Vogt et al. 1986). The phosphorus concentration in this litter, of 0.4‰, was however low compared to values reported for forests worldwide, and immobilisation of phosphorus in the organic layer could be expected (Enright 2001). In addition, a litterbag experiment shows that kauri leaf litter and kauri twigs immobilise phosphorus during the first year of decomposition (Enright and Ogden 1987). Kauri thus contributes to the low availability of phosphorus by the accumulation of phosphorus in the organic layer. Phosphorus is not only bound in organic form, but also in inorganic form, particularly in acidic soils (Vitousek and Farrington 1997). Especially aluminium and iron hydroxides can strongly adsorb phosphate (Jenny 1980). The soils of kauri forests are normally acid and pH- H2O values of 4 have been reported for the A-horizon (Yeates et al. 1981; Jongkind and Buurman 2006). Furthermore, aluminium hydroxides have been found in the clay soils of the Waitakere Ranges (Jongkind and Buurman 2006) and in the clay and sand soils of (Buil 2000). It is therefore likely that the fixation of phosphate in inorganic form, by for instance aluminium hydroxides, is another explanation for the general deficiency of phosphorus in kauri forests. In contrast to phosphorus, nitrogen is only present in low quantities in freshly deposited substrates. Due to nitrogen fixation and atmospheric deposition these quantities normally increase over time (Vitousek and Farrington 1997; Berendse 1998). Still nitrogen is often limiting production in many natural ecosystems (Vitousek 1982; Vitousek and Sanford

87 Chapter 7

Jr 1986; Güsewell 2004; Wassen et al. 2005). This limited availability of nitrogen might be explained by nitrogen being easily lost, by processes like leaching or denitrification, and by nitrogen being strongly retained in organic matter (Vitousek et al. 2002). One place where nitrogen accumulates in kauri forests is the standing biomass of the kauri trees. When growing older, the trunk of a kauri tree gets very thick. The amount of -2 aboveground biomass in a stand of mature kauri trees can be as high as 150 kg m , which is high compared to the biomass in tropical forests of similar structure (Vitousek and Sanford Jr 1986; Silvester and Orchard 1999). However, kauri trees have relatively low concentrations of nitrogen in their leaves compared to other species (Peterson 1962a) and since a large proportion of biomass in mature kauri forests consists of wood, which has a nitrogen concentration of only 0.06% (Silvester 2000), the accumulation of nitrogen in the -2 aboveground biomass of up to 0.1 kg N m (Silvester 2000) is relatively low compared to for instance tropical forests (Vitousek and Sanford Jr 1986). While the accumulation of nitrogen in the standing biomass is relatively low, the annual nutrient requirement of kauri trees appears to have an average value. Two studies estimated the annual amounts of nitrogen returned in litterfall, and found values of 1.6 to 8 g -2 -2 m for four sites (Silvester 2000) and 5.3 g m for another site (Enright 2001). This level of nitrogen return is similar to values reported for warm temperate forests and for subtropical forests (Vogt et al. 1986). The results of my trenching experiment (Table 4.2) also show that the uptake of nitrogen by kauri roots is responsible for the removal of considerable quantities of nitrogen from the forest floor below the kauri crown. The nitrogen which is returned by kauri litterfall is not released quickly but accumulates, as indicated by the high C-N ratios of kauri litterfall (Enright and Ogden 1987). The slow release rate of nitrogen from kauri litter was apparent from the results of my litterbag experiment too (Chapter 6). Nitrogen is retained in large quantities in the organic layer of kauri forests. At the four sites studied by Silvester (Silvester 2000), nitrogen -2 accumulation in the organic layer ranged from 0.05 - 0.65 kg m . This is an extreme -2 accumulation compared to the 0.006-0.03 kg m reported for warm temperate forests or the -2 0.012 kg m reported for subtropical forests worldwide (Vogt et al. 1986). Nitrogen accumulation in the organic layer at another kauri site was found to be much lower, only -2 0.015 kg m (Enright 2001). However, this other site was situated on a steep slope where there was almost no litter building up (Enright 2001), while the four sites of Silvester were situated on flat land (Silvester 2000). Silvester (2000) also estimated the nitrogen accumulation in the mineral soil. For the four sites he studied, the nitrogen accumulation in -2 the top 30 cm of the mineral soil ranged from 0.58 to 0.60 kg m . Altogether there is an extreme accumulation of nitrogen below kauri trees. This is remarkable, since, as described

88 General discussion

above, especially in the soil below the kauri crown nitrogen availability is low (Figure 2.1). Nitrogen availability even decreases when the organic layer gets thicker (Table 2.3). Obviously, the accumulated nitrogen in the organic layer is not available for the uptake by plants (Silvester 2000). For many litter types a reduction of nitrogen mineralisation by tannins has been found (e.g. Chapter 1; Schimel et al. 1996; Bradley et al. 2000; Fierer et al. 2001; Kraus et al. 2003a; Kraus et al. 2004). To study the effects of kauri tannins on nitrogen release in the field, we used two kinds of kauri litter differing 50% in tannin concentration (Chapter 6). While we found no significant difference in nitrogen loss, we did find a stronger increase in nitrogen concentration in the litter with the higher tannin concentration (Figures 6.1 and 6.2, Table 6.3). However, not only the tannin concentrations differed between the two kinds of litter, their initial nitrogen concentrations were different too (Table 6.2). Therefore, the litterbag experiment provided no conclusive evidence in favour or against the hypothesis that kauri tannins affect nitrogen mineralisation in the field (Chapter 6). In a second study, conducted in the laboratory, we added purified kauri tannins to samples of the soil organic layer from under a kauri tree (Chapter 5). Nitrogen availability was reduced by the addition of kauri tannins, both after one month and after one year of incubation (Figures 5.1 and 5.2). During the first month of incubation, kauri tannins reduced nitrogen availability by stimulating immobilisation or by the sequestration of proteins. In the long-term, kauri tannins reduced nitrogen availability by complexing proteins. These results strongly suggest that kauri tannins contribute to the nitrogen accumulation below kauri. An important source for the nitrogen that accumulates below kauri trees seems to be the surrounding angiosperm vegetation. The kauri litter and the litter of the surrounding angiosperm species differ in mineralisation rates. Kauri litter is slow to decompose, releasing few nutrients (see above; Enright and Ogden 1987), while the litter of many angiosperm species decomposes more rapidly, releasing nutrients quicker (Enright and Ogden 1987). Besides, a study in mixed forests of angiosperm and kauri trees (Enright 1999) shows that 55% of the litterfall in these mixed forests was from gymnosperm trees, predominantly kauri, but that only 30-45% of the nutrients in the overall litterfall came from these gymnosperm trees. Kauri litterfall was furthermore concentrated immediately below these trees while the litterfall of the angiosperm trees was more broadly distributed across the forest floor (Enright 2001). By these differences in litter quality and litter distribution more nutrients are transported from the angiosperm trees to the kauri trees than the other way around (Ogden and Stewart 1995; Enright 2001). A second source of nitrogen is the fixation of gaseous nitrogen by bacteria in the forest floor of kauri forests (Silvester and Bennet 1973; Silvester 1978). The fixation appears

89 Chapter 7

to occur especially in the upper fermentation layer and seems primarily limited by the availability of phosphorus. Nitrogen fixation is estimated to give inputs of 0.4-0.7 g nitrogen -2 m per year, and during the lifetime of a kauri tree, of 600-700 years, this might result in -2 inputs of 0.5 kg nitrogen m (Silvester 2000). Kauri forest is not only a mosaic of soil fertility, but is also a mosaic of water availability, which is suggested by the measurements of the soil moisture availability below and outside kauri trees (Chapter 2). During a dry period the organic layer below the kauri crown was found to be dryer than the mineral soil further from the crown (Table 2.2). The trenching treatment of the field experiment (Chapter 4) shows that the uptake of water by kauri roots might be an important factor in explaining the dryness of the soil below kauri trees (Table 4.1). Canopy openness below kauri trees appears to increase with increasing age of the kauri forest and increasing thickness of the organic layer, what was revealed by our study of the site conditions below kauri (Table 2.3). This seems due to the poor growth of many shrub and understory tree species below the old kauri trees. As the kauri tree grows older and thicker, more nutrients accumulate in the organic layer, decreasing the nutrient availability below the tree (Chapter 2). The growth of plants surrounding the trunk probably declines as a result of the low nutrient availability.

Effects on seedlings

Compared to the seedling numbers in tea tree vegetation, I only observed low numbers of kauri seedlings in kauri stands (Table 7.2). Other studies report similar findings (e.g. Mirams 1957; Esler and Astridge 1974; Halkett 1983). The main cause for the death of many young kauri seedlings below kauri appears to be the dryness of the soil (Mirams 1957). This is indicated in my study on seedling distribution below kauri as well, by the positive correlation between the number of small kauri seedlings, their survival, and soil moisture availability (Table 3.2).

Table 7.2. The mean (±SE) number of kauri seedlings (per100 m2) of three different height classes in the kauri and tea tree plots of the studies described in chapters 2 and 3. Column values followed by different letters are significantly different (P<0.05). 0-10 cm 10-30 cm 30-100 cm Kauri plots 17.4 ± 6.5 a 1.8 ± 1.6 a 0.6 ± 0.6 a Tea tree plots 36.8 ± 20.1 a 38.7 ± 14.8 b 17.9 ± 2.8 b

90 General discussion

The observation that the density of kauri seedlings is generally much higher in tea tree vegetation than in stands of mature kauri trees is probably responsible for the common idea (e.g. Cheeseman 1914; Sando 1936; Mirams 1957; Bieleski 1959) that kauri regenerates poorly in mature kauri forests (Ogden et al. 1987). However, a study into the diameter distribution of 25 kauri stands situated throughout northern New Zealand (Ahmed and Ogden 1987) revealed that mature trees of all diameter classes are present in most mature kauri stands. This strongly suggests that kauri regeneration does occur in mature stands. Another -2 study revealed the presence of considerable numbers (4-8 per 100 m ) of seedlings (<2 cm dbh) and sapling (2-10 cm dbh) in many mature kauri stands (Ogden et al. 1987), giving the same indication. During my field work I observed the presence of seedlings (Table 7.2) and young kauri trees in stands of mature kauri trees too. The presence of kauri regeneration in mature stands led to the formulation of a ‘cohort regeneration model’ for kauri (Beveridge 1977; Ogden 1985; Ahmed and Ogden 1987; Ogden et al. 1987; Ogden and Stewart 1995). Before the formulation of this model, a kauri stand was often considered to be a successional stage, that will be replaced by angiosperm or podocarp forest (Mirams 1957; Ecroyd 1982; Ogden et al. 1987). In the cohort regeneration model of Ogden et al., it is expected that, following a large-scale disturbance, like a cyclone (Ogden et al. 1991), fire (Ogden et al. 1998) or landslide (Claessens 2005) kauri is able to spread across the landscape and establish dense regeneration at new sites. Over time, stands of poles or rickers arise and after some hundreds of years, the initial kauri trees start to die. Canopy gaps are formed and, because of the presence of kauri regeneration, the initial cohort of kauri trees is potentially replaced by a second generation (Chapter 1; Ogden 1985; Ogden and Stewart 1995; Enright et al. 1999). My studies of the conditions and seedlings below and outside kauri (Chapters 2, 3, 4) contribute to this theory by shedding light on the regeneration process within a kauri stand. While young kauri seedlings are susceptible to drought (see above, Table 3.2), older kauri seedlings are probably better resistant to drought than many other species (Stephens et al. 1999), which might be explained by the thick and heavy cuticle of kauri leaves (Barton 1982). My study of the seedling distribution in kauri forest suggests that the seedlings of angiosperm tree and shrub species suffer, like young kauri seedlings, from the dryness below the kauri trees (Chapter 3). This is indicated by the positive correlations which I found between soil moisture availability and their survival (Table 3.2). The results of my field experiment suggest as well that both kauri seedlings and angiosperm seedlings suffer from dryness below kauri (Chapter 4). Kauri seedlings in addition, do not require much nutrients for their growth. The low nutrient requirement of kauri seedlings was observed by Peterson (1962a) who grew kauri seedlings on a range of nutrient solutions. My study of the seedlings distribution below

91 Chapter 7

and outside kauri also suggests a low nutrient requirement of kauri seedlings, since I found no correlation between the presence of kauri seedlings and nutrient availability (Table 3.1). By contrast, for angiosperm seedlings I found positive correlations between their survival or presence and the availability of nitrogen and phosphorus (Chapter 3), suggesting that they have higher nutrient requirements than kauri. Furthermore, the relatively open conditions below older kauri trees might favour the growth and establishment of kauri seedlings (Chapter 3). Kauri seedlings need relatively high light intensities for their growth (Table 3.1; Bieleski 1959; Pook 1979; Wright 1993), and my study of the site conditions below kauri (Chapter 2) revealed that, as kauri trees grow older and the organic layer gets thicker, canopy openness increases (Table 2.3). The study of the seedling distribution below and outside kauri (Chapter 3) shows that the survival of angiosperm seedlings is lower below the kauri crown than outside it (Figure 3.1). This suggests that the nutrient poor and dry conditions below kauri decrease the competition by angiosperm seedlings for kauri seedlings. The results of my field experiment lead to the same conclusion (Chapter 4). Nutrient poor and dry conditions resulted in a much larger growth reduction of the mapau (Myrsine australis) seedlings, an angiosperm species, than of kauri seedlings. Apparently kauri seedlings suffer less from the competition by angiosperm seedlings below the kauri crown than outside the kauri crown, where nutrient and water availability are higher (Chapters 3 and 4). For that reason, the site below the kauri crown is an important place for kauri regeneration.

A positive feedback between kauri and the soil

As discussed, kauri trees contribute to the low availability of nitrogen and phosphorus in the soil below their canopy, by accumulating these elements in their organic layer. Kauri is not unique in this regard, since many examples exist of plant species that alter their environment via the litter they produce (e.g. Van Breemen 1993; Berendse 1994a; Binkley 1995; Wardle 2002; Kraus et al. 2003a). My studies on tree seedlings in kauri forest strongly suggest that by lowering nutrient availability, kauri trees lessen the competition from fast growing angiosperm seedlings for kauri seedlings (see discussion above). It appears that, by contributing to the low availability of nutrients below kauri, kauri litter creates an environment where kauri seedlings are better able to compete with angiosperm seedlings. A positive feedback between kauri and the soil, whereby the influence of kauri litter on the soil induces changes in the soil that positively affect the fitness of kauri, therefore seems possible. Other examples of positive feedbacks between plant species and the soil, mediated by the

92 General discussion

plant litter, exist (Chapter 1; Binkley and Giardina 1998), like for instance, the forest mosaic of hemlock (Tsuga canadensis) and sugar maple (Acer saccharum) in the United States (Frelich et al. 1993), or the competition between Ericaceous species and grass species in heathland vegetation in the Netherlands (Berendse 1994b; Berendse 1998). To assess whether soil-plant feedbacks are driven by natural selection, it is not only necessary to test if plants affect their own fitness via soil processes, but also that individuals within a species display genetic variation in the attributes that influence soil properties (Van Breemen and Finzi 1998). For kauri, the tannins in its foliage appear to contribute to the low nitrogen availability in the soil beneath its crown (Chapter 5). Measurements of the tannin concentration in kauri foliage showed that kauri has not an extremely high tannin concentration it its foliage, compared to other species (Table 6.1). It is probably the combination of kauri’s longevity, its ability to occupy certain positions in the landscape for generations of trees and the presence of tannins in kauri foliage, which explains the extreme accumulation of nitrogen in the organic layer below kauri (Chapters 5 and 6). The experiment in which kauri seedlings, raised from seeds of different trees and forest areas, were grown in the same environment, suggests that the variation in the tannin concentration of kauri foliage is partly heritable (Chapter 6). Therefore, natural selection of kauri trees with high tannin concentration is possible, and the feedback between kauri and the soil might be driven by evolution. In conclusion, it seems that the results of my studies confirm the main hypothesis. Kauri tannins appear to have the ability to reduce nitrogen mineralisation in the soil below the kauri crown, thereby creating an environment where kauri seedlings are better able to compete with angiosperm seedlings, which can result in natural selection for kauri trees with high tannin concentration.

93 References

94 References

References

Ahmed, M., Ogden, J. 1987. Population dynamics of the emergent conifer Agathis australis (D. Don) Lindl. (kauri) in New Zealand; I. Population structures and tree growth rates in mature stands. New Zealand Journal of Botany 25: 217-229. Barton, I.L. 1982. An investigation of aspects of the physiology and ecology of kauri (Agathis australis Salisb.). Unpublished MSc-thesis, University of , Hamilton, New Zealand. Barton, I.L., Madgwick, H.A.I. 1987. Response of a kauri stand to fertilizer addition and thinning. New Zealand Forestry 32: 19-21. Beever, J. 1981. A map of the pre-European vegetation of Lower Northland, New Zealand. New Zealand Journal of Botany 19: 105-110. Berendse, F. 1994a. Litter decomposability, a neglected component of plant fitness. Journal of Ecology 82: 187- 190. Berendse, F. 1994b. Competition between plant populations at low and high nutrient supplies. Oikos 71: 253- 260. Berendse, F. 1998. Effects of dominant plant species on soils during succession in nutrient-poor ecosystems. Biogeochemistry 42: 73-88. Berendse, F., Elberse, W.T. 1990. Competition and nutrient availability in heathland and grassland ecosystems. In: Grace, J.B., Tilman, D. (eds). Perspectives on plant competition. Academic Press, pp. 93-116. Berger, T.W., Köllensperger, G., Wimmer, R. 2004. Plant-soil feedback in spruce (Picea abies) and mixed spruce-beech (Fagus sylvatica) stands as indicated by dendrochemistry. Plant and Soil 264: 69-83. Beveridge, A.E. 1975. Kauri forests in the New Hebrides. Philosophical transactions of the Royal Society of London, series B 272: 369-383. Beveridge, A.E. 1977. Notes on the silviculture of kauri. In: Chavasse, C.G.R. (ed). Forestry handbook 1977. New Zealand Institute of Foresters, pp. 125-130. Bieleski, R.L. 1959. Factors affecting growth and distribution of Kauri (Agathis australis Salisb.). Australian Journal of Botany 7: 252-294. Bieleski, R.L. 1979. Kauri Forest. In: New Zealand Trees, . Collins Nature Heritage Series, Auckland and London, pp. 18-28. ISBN 0 00 216949 5. Binkley, D. 1995. The influence of tree species on forest soils: processes and patterns. In: Mead, D.J., Cornforth, I.S. (eds). Proceedings of the Trees and Soil Workshop, Lincoln University, New Zealand, 28 February - 2 March 1994. pp. 1-33. Binkley, D., Giardina, C. 1998. Why do tree species affect soils? The Warp and Woof of tree-soil interactions. Biogeochemistry 42: 89-106. Bloomfield, C. 1953. A study of podzolization. Part II. The mobilization of iron and aluminium by the leaves and bark of Agathis australis (kauri). Journal of Soil Science 4: 17-23. Bond, W.J. 1989. The tortoise and the hare: ecology of angiosperm dominance and gymnosperm persistence. Biological Journal of the Linnean Society 36: 227-249. Bosatta, E., Staaf, H. 1982. The control of nitrogen turn-over in forest litter. Oikos 39: 143-151. Bradley, R.L., Titus, B.D., Preston, C.P. 2000. Changes to mineral N cycling and microbial communities in

black spruce humus after additions of (NH4)2SO4 and condensed tannins extracted from Kalmia angustifolia and balsam fir. Soil Biology and Biochemistry 32: 1227-1240.

95 References

Buil, S.2000. Soil formation below kauri (Agathis australis), New Zealand. Unpublished MSc-thesis, Soil Science Department, Wageningen University, The Netherlands. Burnham, C.P. 1989. Pedological processes and nutrient supply from parent material in tropical soils. In: Proctor, J. (ed). Mineral nutrients in tropical forest and savanna ecosystems. pp. 27-41. Burns, B.R., Leathwick, J.R. 1996. Vegetation - environment relationships at Waipoua Forest, Northland, New Zealand. New Zealand Journal of Botany 34: 79-92. Burns, B.R., Smale, M.C. 1990. Changes in structure and composition over fifteen years in a secondary kauri ( Agathis australis) - tanekaha (Phyllocladus trichomanoides) forest stand, , New Zealand. New Zealand Journal of Botany 28: 141-158. Bradley, R.L., Titus, B.D., Preston, C.P. 2000. Changes to mineral N cycling and microbial communities in

black spruce humus after additions of (NH4)2SO4 and condensed tannins extracted from Kalmia angustifolia and balsam fir. Soil Biology and Biochemistry 32: 1227-1240. Brophy, J.J., Goldsack, R.J., Wu, M.Z., Fookes, C.J.R., Forster, P.I. 2000. The steam volatile oil of Wollemia nobilis and its comparison with other members of Araucariacea (Agathis and Araucaria). Biochemical Systematics and Ecology 28 (2000) 563-578. Chapin, F.S. 1980. The mineral nutrition of wild plants. Annual Review of Ecology Evolution and Systematics 11: 233-260. Cheeseman, T.F. 1914. Illustrations of the New Zealand Flora. Published under the authority of the Government of New Zealand, Government Printer, Wellington, New Zealand. Claessens, L. 2005. Modelling Landslide Dynamics in Forested Landscapes. Addressing landscape evolution, landslide soil redistribution and vegetation patterns in the Waitakere Ranges, west Auckland, New Zealand. PhD-thesis, Wageningen University, The Netherlands. Clark, B.R., Hartley, S.E., Suding, K.N., de Mazancourt, C. 2005. The effect of recycling on plant competitive hierarchies. The American Naturalist 165: 609-622. Clarkson, F.H.1992. Canopy physiology of Agathis australis Salisb. (Kauri). A thesis submitted in partial fulfilment of the requirements for the degree of Master of Science in Biological Sciences. Unpublished MSc-thesis, University of Waikato, Hamilton, New Zealand. Dijkstra, F.A. 2001. Effects of tree species on soil properties in a forest of the northeastern United States. PhD- thesis, Wageningen University, The Netherlands. Eagle, M.K. 2002. Fossil Agathis (Plantae: Araucariaceae) in New Zealand. International Araucariaceae Symposium, Auckland, New Zealand14-17 March 2002, International Dendrology Society. pp. 28-29. Ecroyd, E. 1982. Biological 8. Agathis australis (D. Don) Lindl. (Araucariaceae) Kauri. New Zealand Journal of Botany 20: 17-36. Enright, N.J. 1999. Litterfall dynamics in a mixed conifer-angiosperm forest in northern New Zealand. Journal of Biogeography 26: 149-157. Enright, N.J. 2001. Nutrient accessions in a mixed conifer - angiosperm forest in northern New Zealand. Austral Ecology 26: 618-629. Enright, N.J., Bartlett, R.M., De Freitas, C.R. 1993. Patterns of species composition, recruitment, and growth within canopy gaps in two New Zealand kauri (Agathis australis) forests. New Zealand Journal of Botany 31: 361-373. Enright, N.J., Ogden, J. 1987. Decomposition of litter from common woody species of kauri (Agathis australis Salisb.) forest in northern New Zealand. Australian Journal of Ecology 12: 109-124.

96 References

Enright, N.J., Ogden, J., Riggs, L.S. 1999. Dynamics of forests with Araucariaceae in the western Pacific. Journal of Vegetation Science 10: 793-804. Englund, S.R., O'Brien, J.J., Clark, D.B. 2000. Evaluation of digital and film hemispherical photography and spherical densiometry for measuring forest light environments. Canadian Journal of Forest Research 30: 1999-2005. Esler, A.E., Astridge, S.J. 1974. Tea (Leptospermum) communities of the Waitakere Range, Auckland, New Zealand. New Zealand Journal of Botany 12: 485-501. Field, J.A., Lettinga, G. 1992. Toxicity of tannic compounds to . In: Hemingway, R.W., Laks, P.E. (eds). Plant polyphenols. Synthesis, properties, significance. Plenum Press, New York, pp. 673- 692. Fierer, N., Schimel, J.P., Cates, R.G., Zou, J. 2001. Influence of balsam poplar tannin fractions on carbon and nitrogen dynamics in Alaskan taiga floodplain soils. Soil Biology and Biochemistry 33: 1827-1839. Frelich, L.E., Calcote, R.R., Davis, M.B., Pastor, J. 1993. Patch formation and maintenance in an old-growth Hemlock-hardwood forest. Ecology 74: 513-527. Gifford, E.M., Foster, A.S. 1989. Chapter 17. Coniferophyta. Coniferales, Voltziales, and Cordaitales. In: Gifford, E.M., Foster, A.S. (eds). Morphology and evolution of vascular plants. W. H. Freeman and Company, pp. 401-453. Griffiths, A.J.F., Miller, J.H., Suzuki, D.T., Lewontin, R.C., Gelbart, W.M. 1996. An introduction to genetic analysis. Freemand and Company, New York. Güsewell, S. 2004. N:P ratios in terrestrial plants, variation and functional significance. New Phytologist 164: 243-266. Halkett, J., Sale, E.V. 1986. The world of the Kauri. Reed Methuen, Auckland, New Zealand. Halkett, J.C. 1983. A basis for the management of New Zealand Kauri (Agathis australis (D. Don) Lindl.) forest. New Zealand Journal of Forestry 28: 15-23. Harborne, J.B. 1997. Role of Phenolic Secondary Metabolites in Plants and their Degradation in Nature. In: Cadisch, G., Giller, K.E. (eds). Driven by Nature: Plant Litter Quality and Decomposition. CAB International, pp. 76-74. Hättenschwiler, S., Hagerman, A.E., Vitousek, P.M. 2003. Polyphenols in litter from tropical montane forests across a wide range in soil fertility. Biogeochemistry 64: 129-148. Hättenschwiler, S., Vitousek, P.M. 2000. The role of polyphenols in terrestrial ecosystem nutrient cycling. Trends in Ecology and Evolution 15: 238-243. Hernes, P.J., Benner, R., Cowie, G., Goni, M.A., Bergamaschi, B.A., Hedges, J.I. 2001. Tannin diagenesis in mangrove leaves from a tropical estuary: a novel molecular approach. Geochimica et Cosmochimica Acta 65: 3109-3122. Hill, R.S., Scriven, L.J. 1998. The fossil record of conifers in Australia. Volume 48, Ferns, Gymnosperms and Allied Groups. CSIRO Publishing, pp. 527-537. Hobbie, S.E. 1992. Effects of Plant Species on Nutrient Cycling. Trends in Ecology and Evolution 7: 336-339. Hommel, P.W.F.M., Spek, T. De Waal, R.W. 2002. Boomsoort, strooiselkwaliteit en ondergroei in loofbossen op verzuringsgevoelige bodem. Een verkennend literatuur- en veldonderzoek. Alterra-rapport 509, Alterra Research Instituut voor de Groene Ruimte, Wageningen, The Netherlands. Jenny, H. 1980. The Soil Resource, Origin and Behavior. Springer-Verlag. New York, 377 pp. Jongkind, A.G., Buurman, P. 2006. The effect of kauri (Agathis australis) on grain size distribution and clay mineralogy of andesitic soils in the Waitakere Ranges, New Zealand. Geoderma 134: 171-186.

97 References

Koerselman, W., Meuleman, A.F.M. 1996. The vegetation N:P ratio: a new tool to detect the nature of nutrient limitation. Journal of Applied Ecology 33: 1441-1450. Kooijman, A.M., Besse, M. 2002. The higher availability of N and P in lime-poor than in lime-rich coastal dunes in the Netherlands. Journal of Ecology 90: 394-403. Kraus, T.E.C., Dahlgren, R.A., Zasoski, R.J. 2003a. Tannins in nutrient dynamics of forest ecosystems. A review. Plant and Soil 256: 41-66. Kraus, T.E.C., Yu, Z., Preston, C.M., Dahlgren, R.A., Zasoski, R.J. 2003b. Linking chemical reactivity and protein precipitation to structural characteristics of foliar tannins. Journal of Chemical Ecology 29: 703- 730. Kraus, T.E.C., Zasoski, R.J., Dahlgren, R.A., Horwarth, W.R., Preston, C.M. 2004. Carbon and nitrogen dynamics in a forest soil amended with purified tannins from different plant species. Soil Biology and Biochemistry 36: 309-321. Kuiters, A.T. 1990. Role of phenolic substances from decomposing forest litter in plant-soil interactions. Acta Botanica Neerlandica 39: 329-348. Lorenz, K., Preston, C.M. 2002. Characterization of high-tannin fractions from humus by Carbon-13 Cross- Polarization and Magic-Angle Spinning Nuclear Magnetic Resonance. Journal of Environmental Quality 31: 431-436. Lovis, J.D. 2001. Book review of the Flora of Australia Volume 48. Ferns, Gymnosperms and Allied Groups. New Zealand Journal of Botany 39: 553-556. Madgwick, H.A.I., Oliver, G., Holten-Anderson, P. 1982. Above-ground biomass, nutrients, and energy content of trees in a second-growth stand of Agathis australis. New Zealand Journal of Forestry Science 12: 3- 6. Mc Kelvey, P.J., Nicholls, J.L. 1959. The indigenous forest types of North Auckland. New Zealand Journal of Forestry 8: 29-45. Mc Kinnon, A.D. 1945. Natural regeneration of kauri. New Zealand Journal of Forestry 5: 133-137. Meletiou-Christou, M.S., Rhizopoulou, S., Diamantoglou, S. 1994. Seasonal changes of carbohydrates, lipids and nitrogen content in sun and shade leaves from four Mediterranean evergreen sclerophylls. Environmental and Experimental Botany 34: 129-140. Mirams, R.V. 1957. Aspects of the natural regeneration of the Kauri (Agathis australis Salisb.). Transactions of the Royal Society of New Zealand 84: 661-680. Molloy, L. 1993. Soils in the New Zealand Landscape, the living mantle. New Zealand Society of Soil Science. Canterbury, New Zealand. Northup, R.R., Dahlgren, R.A., Mc Coll, J.G. 1998. Polyphenols as regulators of plant-litter-soil interactions in northern California's pygmy forest: a positive feedback? Biogeochemistry 42: 189-220. Northup, R.R., Yu, Z., Dahlgren, R.A., Vogt, K.A. 1995. Polyphenol control of nitrogen release from pine litter. Nature 377: 227-229. Ogden, J. 1985. An introduction to plant demography with special reference to New Zealand trees. New Zealand Journal of Botany 23: 751-772. Ogden, J., Basher, L., Mc Glone, M. 1998. Fire, Forest Regeneration and Links with Early Human Habitation: Evidence from New Zealand. Annals of Botany 81: 687-696. Ogden, J., Fordham, R.A., Pilkington, S., Serra, R.G. 1991. Forest gap formation and closure along an altitudinal gradient n Tongariro National Park, New Zealand. Journal of Vegetation Science 2: 165-172.

98 References

Ogden, J., Stewart, G.H. 1995. Community Dynamics of the New Zealand Conifers. In: Enright, N.J., Hill, R.S. (eds). Ecology of the Southern Conifers. Melbourne University Press, pp. 81-119. Ogden, J., Wardle, G.M., Ahmed, M. 1987. Population dynamics of the emergent conifer Agathis australis (D. Don) Lindl. (kauri) in New Zealand; II. Seedling population sizes and gap-phase regeneration. New Zealand Journal of Botany 25: 231-242. Pearcy, R.W., Ehleringer, J., Mooney, H.A., Rundel, P.W. 1989. Plant physiological ecology: field methods and instrumentation. Chapman and Hall. New York. Peterson, P.J. 1961. Variation in the mineral content of kauri (Agathis australis Salisb.) leaves with respect to leaf age, leaf position, and tree age. New Zealand Journal of Science 4: 669-678. Peterson, P.J. 1962a. Mineral nutrition of Agathis australis Salisb., the kauri. Part I. Effects of deficiencies of essential elements of the growth and foliar mineral composition of seedlings. New Zealand Journal of Science 5: 141-164. Peterson, P.J. 1962b. Mineral nutrition of Agathis australis, Salisb., the kauri. Part 2. An extension to forest grown seedlings and including analyses of tree foliage. New Zealand Journal of Science 5: 221-236. Podger, F.D., Newhook, F.J. 1971. in Indigenous Plant Communities in New Zealand. New Zealand Journal of Botany 9: 625-638. Pook, E.W. 1978. Population and growth characteristics of tanekaha in Leptospermum scrub and secondary forest of the Waitakere Ranges, Auckland, New Zealand. New Zealand Journal of Botany 16: 227-234. Pook, E.W. 1979. Seedling growth in Tanekaha (Phyllocladus trichomanoides): effects of shade and other seedling species. New Zealand Journal of Forestry Science 9: 193-200. Prescott, C.E. 2002. The influence of the forest canopy on nutrient cycling. Tree Physiology 22: 1193-1200. Reed, A.H. 1954. The story of the kauri. Reed, Wellington, New Zealand. Sale, E. V. 1978. Quest for the Kauri, forest giants and where to find them. Reed, Wellington, New Zealand. 230 pp. Sando, C.T. 1936. Notes on Agathis australis. New Zealand Journal of Forestry 4: 16-21. Schimel, J.P., Van Cleve, K., Cates, R.G., Clausen, T.P., Reichardt, P.B. 1996. Effects of balsam poplar (Populus balsamifera) tannins and low molecular weight phenolics on microbial activity in taiga floodplain soil: implications for changes in N cycling during succession. Canadian Journal of Botany 74: 84-90. Schofield, J.A., Hagerman, A.E., Harold, A. 1998. Loss of tannins and other phenolics from willow leaf litter. Journal of Chemical Ecology 24: 1409-1421. Schultz, J.C., Hunter, M.D., Appel, H.M. 1992. Antimicrobial activity of polyphenols mediates plant-herbivore interactions. In: Hemingway, R.W., Laks, P.E. (eds). Plant Polyphenols. Synthesis, properties, significance. Plenum Press, New York, pp. 621-637. Silvester, W.B. 1978. Nitrogen Fixation and Mineralisation in Kauri (Agathis australis) Forest in New Zealand. In: Loutit, M.W., Miles, J.A.R. (eds). Microbial Ecology. Springer-Verlag, Berlin, Heidelberg, New York, pp. 139-143. Silvester, W.B. 2000. Nitrogen cycling in kauri (Agathis australis) forest: an example of extreme accumulation, fixation and immobilisation of nitrogen. New Zealand Journal of Botany 38: 205-220. Silvester, W.B., Bennet, K.J. 1973. Acetylene reduction by roots and associated soil of New Zealand conifers. Soil Biology and Biochemistry 5: 171-179.

99 References

Silvester, W.B., Orchard, T.A. 1999. The biology of kauri (Agathis australis Salisb.) in New Zealand. I Production, biomass, carbon storage and litterfall in four forest remnants. New Zealand Journal of Botany 37: 553-571. Stephens, D.W., Silvester, W.B., Burns, B.R. 1999. Differences in water-use efficiency between Agathis australis and Dacrycarpus dacrydioides are genetically, not environmentally, determined. New Zealand Journal of Botany 37: 361-367. Swindale, L.D. 1957. The effect of kauri vegetation upon the development of soils from rhyolite and olivine basalt. New Zealand Soil News 5: 115-118. Temminghof, E.J.M., Houba, V.J.G., Van Vark, W., Gaikhorst, G.A. 2000. Soil and plant analysis procedures. Part 3: plant analysis procedures. Wageningen University, sub-department of soil quality, The Netherlands. Ter Steege, H. 1996. Winphot 5; a programme to analyze vegetation indices, light and light quality from hemisperical photographs. Tropenbos Guyana Reports 95-2, Tropenbos Guyana Programme, Georgetown Guyana. Thomas, B.R. 1969. Kauri - Modern and Fossil. In: Englinton, G., Murphy, M.T.J. (eds). Organic Geochemistry, Methods and Results. Springer-Verlag, Berlin, Heidelberg, New York, pp. 599-618. Van Breemen, N. 1993. Soils as biotic constructs favouring net primary productivity. Geoderma 57: 183-211. Van Breemen, N., Finzi, A.C. 1998. Plant-soil interactions: ecological aspects and evolutionary implications. Biogeochemistry 42: 1-19. Van Breemen, N. Scott, N. 1998. Plant /soil interactions: Weahtering processes in New Zealand podzols. Project Report, Landcare Research New Zealand. Van Genderen, H.; Schoonhoven, L.M., Fuchs, A. 1996. Chemisch-ecologische flora van Nederland en België. Een inleiding over aard en ecologische betekenis van secundaire plantestoffen. KNNV Uitgeverij, 298 pp. Vitousek, P. 1982. Nutrient cycling and nutrient use efficiency. The American Naturalist 119 (4): 553-572. Vitousek, P.M., Farrington, H. 1997. Nutrient limitation and soil development: experimental test of a biogeochemical theory. Biogeochemistry 37: 63-75. Vitousek, P.M., Hättenschwiler, S., Olander, L., Allison, S. 2002. Nitrogen and Nature. Ambio 31: 97-101. Vitousek, P.M., Sanford Jr, R.L. 1986. Nutrient cycling in moist tropical forest. Annual Review of Ecology Evolution and Systematics 17: 137-167. Vogt, K.A., Grier, C.C., Vogt, D.J. 1986. Production, Turnover, and Nutrient Dynamics of Above- and Belowground Detritus of World Forests. Advances in Ecological Research 15: 303-377. Wardle, P. 1991. Vegetation of New Zealand. Cambridge University Press. ISBN 0 521 25873 1. Wardle, D. A. 2002. Communities and Ecosystems: linking the aboveground and belowground components. Monographs in population biology. Princeton University Press. ISBN 0 691 07487 9. Wassen, M.J., Venterink, H.O., Lapshina, E., Tanneberger, F. 2005. Endangered plants persist under phosphorus limitation. Nature 437: 547-550. Waterman, P.G., Mole, S. 1994. Analysis of Phenolic Plant Metabolites. Blackwell Scientific Publications, Oxford. The Methods in Ecology Series, 238 pp. Wells, N., Northey, R.D. 1985. Strengths of a densipan, humus-pan and clay-pan in a spodosol developed under kauri (Agathis australis) and the implication for soil classification. Geoderma 35: 1-13. Whitmore, T.C., Page, C.N. 1980. Evolutionary implications of distribution and ecology of the tropical conifer Agathis. New Phytologist 84: 407-416.

100 References

Wright, S.D.T. 1993. The comparative regeneration ecology of three tree species in Northern New Zealand. Unpublished MSc-thesis, University of Auckland, New Zealand. Yeates, G.W., Cox, J.E., Ross, D.J., Stout, J.D. 1981. Biology and biochemistry of the soil profile below a kauri (Agathis australis) tree. New Zealand Soil Bureau Scientific Report 51, Department of Scientific and Industrial Research, 34 pp. Yu, Z., Dahlgren, R.A. 2000. Evaluation of methods for measuring polyphenols in conifer foliage. Journal of Chemical Ecology 26: 2119-2140. Yu, Z.S., Northup, R.R., Dahlgren, R.A. 1994. Determination of dissolved organic nitrogen using persulfate oxidation and conductimetric quantification of nitrate nitrogen. Communication in Soil Science and Plant Analysis 25: 3161-3169. Zeller, B., Coline-Belgrand, M., Dambrine, E., Martin, F., Bottner, P. 2000. Decomposition of 15N-labelled beech litter and fate of nitrogen derived from litter in a beech forest. Oecologia 123: 550-559.

101 Samenvatting

Samenvatting

Bomen hebben een grote invloed op bodemprocessen, zoals op de mineralisatie van nutriënten en de afbraak van organische stof. Niet alle boomsoorten hebben dezelfde invloed op de bodem, met als gevolg dat bodemeigenschappen kunnen gaan verschillen tussen boomsoorten die oorspronkelijk op dezelfde bodem groeiden. De verschillende effecten die de boomsoorten hebben op de bodem kunnen ook leiden tot verschillen in de vegetatie onder de bomen. Maar het zou ook de overleving en vestiging van de boomsoorten zelf kunnen beïnvloeden. Mogelijk ontstaat er een positieve terugkoppeling, waarbij een boomsoort bodemeigenschappen verandert die gunstig zijn voor de overleving van die boomsoort. Polyfenolen zouden een rol kunnen spelen bij zo’n terugkoppeling. Polyfenolen, en dan met name de grotere polyfenolmoleculen zoals tanninen of looistoffen, kunnen eiwitten binden door de vorming van waterstofbruggen. Op die manier kunnen ze de afbraak van eiwitten tegengaan. Ze kunnen door te binden aan de enzymen van herbivoren ook de vraat door herbivoren remmen. Men gaat er dan ook vanuit dat hoge polyfenolconcentraties in bladeren, die bladeren beschermen tegen vraat van herbivoren. Wanneer bladeren met een hoge polyfenolconcentratie afvallen, hebben ze echter ook een effect op de processen in de bodem. In de bodem kunnen polyfenolen de mineralisatie van stikstof remmen. Daardoor zouden ze de bodem minder geschikt kunnen maken voor plantensoorten die relatief veel stikstof nodig hebben. Omdat er nog maar weinig bekend is over de gevolgen van bodemveranderingen door planten op de overleving van die planten, en omdat het niet duidelijk is of tanninen hierbij een rol spelen werd het onderzoeksproject ‘Podzolisatie onder kauri’ gestart. Met dit project werd onderzoek gedaan naar het effect dat de Nieuw-Zeelandse boomsoort kauri (Agathis australis) heeft op de bodem en daarmee op de vegetatie. Kauri is om verschillende redenen een interessante boomsoort om onderzoek te doen naar de terugkoppeling tussen bodem en boom en naar de rol van tanninen daarbij. Kauri komt van nature voor op het Noord-eiland van Nieuw-Zeeland. Kauribomen zijn groot en worden gemiddeld 3 meter dik en 30 meter hoog. Ze worden normaal 600-700 jaar oud. Onder volwassen bomen ontstaat een dikke strooisellaag, die bij de stam meer dan 2 meter dik 2 kan zijn. De massa van deze laag kan oplopen tot 55 kg/m . In deze laag is veel stikstof 2 vastgelegd, tot 0,65 kg/ m , dat niet door planten op te nemen is. Het lijkt er verder op dat de dikke strooisellaag zorgt voor een snellere verwering van de bodem onder kauribomen. Kauribos bestaat uit een mozaïek van plekken met kauribomen en plekken die gedomineerd worden door andere boomsoorten. Wanneer er bos verdwijnt, door bijvoorbeeld een brand of een storm, vestigen zich op de open plekken ‘theebomen’ (Leptospermum scoparium en Kunzea ericoides). Vooral onder deze ‘theebomen’ komen veel kaurizaailingen

102 Samenvatting

voor en op den duur zullen deze kaurizaailingen de ‘theebomen’ overgroeien. Er ontstaat dan een groep van volwassen kauribomen van ongeveer dezelfde leeftijd. Onder zo’n groep kauribomen ontwikkelt zich een typische vegetatie waarbij de soorten ‘kauri grass’ (Astelia trinveria) en ‘cutty grass’ (Gahnia xanthocarpa) de kruid- en struiklaag domineren. Ook zijn er kaurizaailingen aanwezig en wanneer de eerste volwassen kauribomen doodgaan worden ze mogelijk opgevolgd door een volgende generatie kauribomen. Vooral op die plekken waar de bodem arm is aan voedingsstoffen worden oude kauribomen opgevolgd door jongere kauribomen. Men vermoedt dat de kaurizaailingen vooral op een arme bodem in staat zijn succesvol te concurreren met loofboomzaailingen.

De hypothese die ik tijdens mijn promotieonderzoek bestudeerde luidt: kauri-tanninen vertragen de stikstofmineralisatie in de bodem onder kauribomen waardoor er een milieu ontstaat waarin kaurizaailingen beter in staat zijn te concurreren met loofboomzaailingen, wat resulteert in natuurlijke selectie van kauribomen met een hoog tanninegehalte.

We hebben verschillende onderzoeken uitgevoerd om deze hypothese te toetsen. In hoofdstuk 2 wordt een studie beschreven naar de milieucondities onder en net buiten de kroon van kauribomen. De belangrijkste vraag was of de beschikbaarheid aan nutriënten lager is onder de kauri dan net buiten de kroon van een kauriboom. Voor deze studie hebben we de nutriëntenbeschikbaarheid, het bodemvochtgehalte en het licht gemeten in proefvlakken onder en naast de kronen van vijf kauribomen. De beschikbaarheid van nutriënten werd niet direct gemeten, maar werd bepaald door de nutriëntenconcentraties in het blad van een algemeen voorkomende struiksoort te meten. De nutriëntenbeschikbaarheid blijkt inderdaad lager onder de kauribomen dan net buiten de kauribomen. Daarnaast was, tijdens een droge periode, de organische bodem rond de kauribomen droger dan de minerale bodem iets verder bij de kauribomen vandaan. In Hoofdstuk 3 bestuderen we de vraag of de omstandigheden onder kauri gunstig zijn voor kaurizaailingen. Onder en rond dezelfde vijf kauribomen als gebruikt bij de voorgaande studie (Hoofdstuk 2) heb ik alle boomzaailingen geteld. Door deze zaailingaantallen te combineren met de gegevens over de milieucondities (uit Hoofdstuk 2) kon ik uitzoeken welke factoren belangrijk zijn voor de zaailingen van de verschillende boomsoorten. Het lijkt erop dat de lichte omstandigheden onder kauribomen gunstig zijn voor de vestiging van kaurizaailingen. De lage nutriëntenbeschikbaarheid onder kauribomen lijkt daarnaast ongunstig voor de overleving van loofboomzaailingen maar niet voor de overleving van kaurizaailingen.

103 Samenvatting

In Hoofdstuk 4 wordt een veldexperiment beschreven waarmee onderzocht werd welke factoren de groei van zaailingen onder kauribomen beperken. In dit experiment hebben we onder kauribomen, in proefvlakken van één bij één meter, zaailingen van kauri en van de loofboomsoort mapau (Myrsine australis) geplant. Aan de zaailingen werd één van de volgende behandelingen toegediend: het verwijderen van de kruidlaag, het verwijderen van de strooisellaag, het toedienen van meststoffen of het uitschakelen van wortelconcurrentie door het ingraven van plasticfolie in de bodem rond het proefvlak. Het blijkt dat vooral wateropname door boomwortels, of de combinatie van water- en nutriëntenopname, de groei van zaailingen onder kauri beperkt. De opname van water en nutriënten beperkte de mapauzaailingen meer in hun groei dan kaurizaailingen, wat gunstig zou kunnen zijn voor de overleving van kaurizaailingen. Hoofdstuk 5 beschrijft een laboratoriumexperiment waarmee we onderzochten of kauri-tannine de stikstofmineralisatie in de bodem kan vertragen. We extraheerden kauri- tannine uit kauriblad en de tannine werd vervolgens in petrischaaltjes toegevoegd aan humusrijke grond uit kauribos. Gedurende de eerste maand zorgden de tannine voor een lagere stikstofbeschikbaarheid in het bodemmateriaal door eiwitbinding of doordat de tannine dienden als voedingsbron voor micro-organismen, waardoor de populatie van deze micro- organismen toenam. Op de langere termijn zorgden de tannine voor een afname van de stikstofbeschikbaarheid in het bodemmateriaal door eiwitbinding. In Hoofdstuk 6 worden drie andere studies beschreven die zich op tannine richtten. Ten eerste hebben we het tanninegehalte in kauriblad vergeleken met dat in tien andere plantensoorten die algemeen voorkomen in kauribos. Vergeleken met deze soorten heeft kauri een hoog tanninegehalte. Uit literatuurgegevens over planten uit andere delen van de wereld blijkt echter dat het tanninegehalte van kauri niet extreem hoog is. De tweede studie beschrijft een experiment dat we uitvoerden om de verschillen in afbreekbaarheid en stikstofmineralisatiesnelheid van kauriblad met een hoog en kauriblad met een laag tanninegehalte te meten. De stikstofconcentratie nam sneller toe in het blad met de hoge tannineconcentratie. Maar omdat aan het begin van het experiment behalve de tannineconcentraties ook het stikstofgehalte van de twee bladtypen verschilden, kon ik op basis van dit resultaat geen definitieve conclusies trekken over het effect van tannine op de stikstofmineralisatiesnelheid. In de derde studie heb ik onderzocht of er genetische variatie is in het tanninegehalte in kauriblad. Daartoe heb ik kaurizaden verzameld onder verschillende bomen in drie verschillende bosgebieden en uit deze zaden heb ik zaailingen opgekweekt in een kas. Uit de vergelijking van de tannineconcentraties van deze zaailingen blijkt dat er inderdaad genetische variatie is in het tanninegehalte.

104 Samenvatting

De discussie en conclusie van mijn proefschrift staan in Hoofdstuk 7. Uit mijn onderzoek blijkt dat stikstof en fosfor schaars zijn onder kauri doordat deze voedingsstoffen ophopen in de strooisellaag, waar ze niet beschikbaar zijn voor planten. De tannine in kauristrooisel lijkt bij te dragen aan de ophoping van stikstof. Kauri heeft geen extreem hoog tanninegehalte. Dat er zich zoveel stikstof kan ophopen in de strooisellaag onder kauribomen, lijkt te komen door de combinatie van enerzijds de lange levensduur van kauri en anderzijds het feit dat op bepaalde posities in het landschap vaak generaties achtereen kauribomen staan. Het onderzoek naar het voorkomen van zaailingen onder kauribomen wijst erop dat de concurrentie die kaurizaailingen ondervinden van loofboomzaailingen wordt verminderd door de geringe beschikbaarheid van stikstof onder kauribomen. De lage stikstofbeschikbaarheid in de bodem onder kauribomen, mede veroorzaakt door kauri-tanninen, heeft dus een gunstige invloed op de overleving van kauribomen. Doordat er genetische variatie is in het tanninegehalte is natuurlijke selectie mogelijk ten gunste van kauribomen met een hoog tanninegehalte. Al met al bevestigen de resultaten van mijn onderzoek de centrale hypothese: het lijkt mogelijk dat er natuurlijke selectie plaatsvindt van kauribomen met een hoog tanninegehalte, via het effect dat deze tanninen hebben op de stikstofbeschikbaarheid in de bodem en het effect dat ze daarmee hebben op de concurrentie tussen zaailingen van verschillende boomsoorten.

105 Summary

Summary

Trees have important impacts upon soil processes such as nutrient mineralisation and organic matter dynamics. The effects trees have upon the soil differ among species and as a result different tree species growing on the same site can lead to different soil characteristics. The different effects species have upon the soil can also lead to differences in the understory vegetation and might affect the fitness of the individual trees themselves. A positive feedback can develop between the trees and the soil, whereby the trees affect soil properties and the changes in the soil properties have positive effects on the fitness of the trees. Polyphenols might play an important role in positive feedbacks between trees and soil. Polyphenols, especially large molecules like tannins, can complex proteins through hydrogen bonding. Thereby they can prevent the breakdown of plant proteins by fungi and bacteria and inhibit essential digestive enzymes of herbivores. It is assumed that high leaf polyphenol concentrations reduce biomass losses by pathogens and herbivores. However, a high polyphenol concentration can affect the soil processes too. High polyphenol levels in plant litter might reduce nitrogen mineralisation in the soil, making the soil less favourable for competing species that need higher levels of available nitrogen. Because the effects of plant induced soil changes on the plant’s fitness have not been studied extensively and the role of tannins in such feedback is still unclear, a research programme, titled ‘Podzolisation under kauri, for better or worse?’, was started. In this programme, the effects of the New Zealand kauri tree (Agathis australis) on soil processes and vegetation was studied. For several reasons, kauri is an interesting species for a study of positive feedbacks between tree species and the soil mediated by tannins. Kauri occurs naturally at the North Island of New Zealand. It is a large tree, up to 30 m or rarely up to 60 m tall, with a columnar trunk up to 3 m, exceptionally up to 7 m, in diameter. The normal attainable age of kauri is 600–700 years. Under mature kauri trees very thick organic layers may develop, reaching 2 2 meters or more in depth near the stem, with a biomass of up to 55 kg/m accumulated in this 2 layer. This organic layer contains large amounts of nitrogen, up to 0.65 kg N/m , that seem inaccessible to plants. It appears that this organic layer leads to increased weathering of the underlying soil. Kauri forest itself is a mosaic of vegetation patches consisting of patches dominated by kauri trees and patches dominated by other species. On places where the forest has disappeared, by for instance fire or a cyclone, patches of tea tree vegetation (Leptospermum scoparium and Kunzea ericoides) establish. It is there that many kauri seedlings can be found. These seedlings will eventually replace the tea tree and develop into a relatively even aged stand of mature kauri. Mature kauri stands typically have a herb and shrub layer dominated by

106 Summary

species like kauri grass (Astelia trinervia) and cutty grass (Gahnia xanthocarpa). Kauri seedlings are also present below the mature kauri trees and when the initial cohort of kauri dies a second cohort potentially replaces it. Subsequent cohorts of kauri become progressively restricted to sites of low soil fertility, where it is expected that kauri seedlings can compete most successfully with angiosperm seedlings.

The main hypothesis which I studied in this thesis was: kauri tannins can reduce nitrogen mineralisation in the soil below the crown of a kauri tree and thus create an environment where kauri seedlings are better able to compete with angiosperm seedlings, resulting in natural selection for kauri trees with high tannin concentrations.

To test the hypothesis, I conducted a number of studies. In Chapter 2 I describe a study of the site conditions below and outside the crown of kauri trees. The main question was if nutrient availability is lower below kauri than just outside the kauri crown. In plots under and outside the crown of five mature kauri trees, I measured nutrient availability, soil moisture availability and light intensity. Nutrient availability was not directly measured, but instead I measured the nutrient concentration in the leaves of a common shrub, which was present in each plot. The results show that near the trunk of a kauri tree the nutrient availability is indeed lower than outside the crown. During a dry period, the organic layer surrounding the kauri trunk was dryer than the mineral soil at greater distance from the kauri trunk. Chapter 3 deals with the question whether or not the conditions below the kauri crown are favourable for kauri seedlings. Below and outside the crowns of the same five mature kauri trees that were used in the previous study (Chapter 2), I counted the number of tree seedlings. By linking the data of the site conditions with the seedling numbers, I investigated which site conditions are important for kauri seedlings and its competitors. The results indicate that below kauri the establishment of kauri seedlings is favoured by the open canopy and the resulting high light intensities. The low nutrient availability under kauri appears to be unfavourable to the survival of angiosperm seedlings but not to the survival of kauri seedlings. In Chapter 4 a field experiment is described that focuses on the question: ‘which factors limit the growth of seedlings below the crown of kauri trees?’. In plots of one by one meter, I planted kauri seedlings and seedlings of the angiosperm species mapau (Myrsine australis) and applied the following treatments: removal of herbs, removal of litter, removal of nutrient limitation by fertilization, and elimination of root competition of mature kauri trees by trenching. The results show that below kauri trees the water uptake by tree roots, or the

107 Summary

combination of water and nutrient uptake, hampers the growth of kauri as well as mapau seedlings. The seedlings of mapau are hampered more than the kauri seedlings, which might result in an increased relative fitness of kauri. In Chapter 5 I describe a study conducted in the laboratory to test the hypothesis that kauri tannins can lower nitrogen mineralisation in the soil. Tannins were extracted from kauri leaves and leaves of two other common New Zealand tree species. These tannins were added to soil organic material and this mixture was incubated in Petri-dishes. After one month and after one year the nitrogen mineralisation and the decomposition of the material were determined. The results suggest that during the first month of incubation the added tannins reduced nitrogen availability by sequestering proteins or by stimulating nitrogen immobilisation. In the long-term, the reduced nitrogen release which was found following tannin addition seems attributable to the complexation of proteins by tannins. In Chapter 6 three other studies focussing on tannins are presented. In the first study I investigated if the tannin concentration in kauri foliage is high compared to the concentration in foliage of other New Zealand plant species. I collected foliage of ten plant species that are common in kauri forest and measured their tannin concentrations. The results confirm that the tannin concentration in kauri foliage is indeed higher than in foliage of most of the other species, but it is not extremely high compared to values for species elsewhere in the world mentioned in the literature. The second study focuses on the effect of tannins on nitrogen mineralisation. In a litterbag experiment, I compared the decomposition and mineralisation rate between kauri foliage with low and kauri foliage with high tannin concentrations. I found no difference in nitrogen loss but there was a difference in the increase of the nitrogen concentration. However, not only the initial tannin concentration differed between the two kinds of litter but also their initial nitrogen concentration, and therefore the experiment provided no conclusive evidence in favour or against an effect of kauri tannins on nitrogen mineralisation. In a third study, I tested if there is genetic variation in the tannin concentration of kauri foliage. To do so, in three forest areas I collected seeds of kauri, raised seedlings from them in a greenhouse, and compared their tannin concentrations. The results of this study suggest that the variation in tannin concentration in kauri foliage is partly genetic. The discussion and conclusion of my thesis are presented in Chapter 7. The results of my studies suggest that below kauri trees nitrogen and phosphorus availability are low because these elements accumulate in the organic layer. The results further show that tannins of kauri contribute to the accumulation of nitrogen. Since kauri has not an extremely high tannin concentration in its foliage, it is probably the combination of the presence of tannin in kauri litter, the longevity of kauri trees and the ability of kauri to occupy the same position in

108 Summary

the landscape for generations of trees that explains the extreme accumulation of nitrogen in the organic layer below kauri. The studies of tree seedlings in kauri forest suggest that below kauri the reduced nutrient availability lessens the competition from fast growing angiosperm seedlings for kauri seedlings. By changing the soil, via its litter, kauri thus improves its own fitness. Since the study with kauri seedlings in the greenhouse indicates that genetic variation in the tannin concentration of kauri foliage is present, natural selection of kauri individuals with high tannin concentration seems possible. Therefore I conclude that the results of my studies confirm the main hypothesis: it seems possible that natural selection for kauri trees with high tannin concentration occurs, via the effects of these tannins upon nitrogen mineralisation and thereby on the competition between seedlings.

109 Curriculum vitae

Curriculum vitae

Eric Verkaik werd geboren op twintig november 1972 in Nederhorst den Berg. Hier doorliep hij met veel plezier de lagere school, waarna hij met veel minder plezier het grote Comenius College te Hilversum bezocht. Hij haalde daar in 1991 zijn VWO-diploma, waarna hij aan de HTS in Utrecht de net opgestarte opleiding milieukunde ging volgen. Hoewel hij de propedeuse met goed gevolg afrondde besloot hij toch van opleiding te veranderen omdat er bij deze opleiding vooral technieken werden aangeleerd, maar wat minder over het waarom ervan werd verteld. Het werd uiteindelijk de studie Bosbouw aan de toenmalige Landbouwuniversiteit in Wageningen. Hij koos voor de afstudeerrichting Bosteelt en Bosecologie. Tijdens zijn eerste afstudeervak deed hij onderzoek naar het effect van een verhoogde CO2-concentratie op twee plantensoorten rond een natuurlijke CO2-bron in Italië. Zijn tweede afstudeervak betrof het ontwerpen van een meer natuurlijk bosteeltsysteem, voor een hardhoutooibos langs de Donau in Hongarije. In 1997 studeerde hij met lof af, waarna hij werkzaam was bij de vakgroep Bosbouw. Daar deed hij literatuuronderzoek naar de rol van twee boomsoorten voor mensen in Burkina Faso. Vervolgens was hij werkzaam bij het onderzoeksinstituut Alterra, waar hij onder andere onderzoek deed naar het effect van een meer natuurlijk bosbeheer op de houtproductie in Scandinavië. Hierna werkte hij tot oktober 1999 bij het Europese Bosinstituut in Finland. Het werk betrof de simulatie van de koolstofvastlegging in de bossen van Rusland en Europa. Omdat hij tijdens dit modelleren dikwijls het gevoel had de processen in dat bos nog maar nauwelijks te kennen, bleef het meer praktische bosonderzoek hem trekken. In het jaar 2000 volgde hij nog een aantal chemische en analytische vakken bij Wageningen Universiteit en verder was hij dat jaar werkzaam als student-assistent bij de vakgroep Bosbouw. In september 2000 kon hij aan de slag als AIO bij de leerstoelgroep Natuurbeheer en Plantenecologie van Wageningen Universiteit om bosecologisch onderzoek te doen in kauribos in Nieuw-Zeeland. De resultaten van dat onderzoek zijn vastgelegd in dit proefschrift.

110 Curriculum vitae

List of publications

Verkaik, E., Bartelink, H.H. 1997. Effects of long-term elevated CO2 on foliage characteristics of Quercus ilex L. and Juniperus communis L. In: Mohren, G.M.J. et al. (eds). Impact of Climate Change on Tree Physiology and Forest Ecosystems. Proceedings of an International Conference, 23-26 November, 1996, Wageningen, The Netherlands, pp. 239-242. ISBN 0-7923-4921-0.

Nabuurs, G.J., Dolman, A.J., Verkaik, E., Whitmore, A.P., Daamen, W.P., Oenema, O., Kabat, P., Mohren, G.M.J. 1999. Resolving Issues on Terrestrial Biospheric Sinks in the Kyoto Protocol. Dutch National Research Programme on Global Air Pollution and Climate Change, report 410 200 030, 101 pp.

Nabuurs, G.J., Verkaik, E. 1998. How much forest do you have under the Kyoto Protocol? Change 44:12-15.

Nabuurs, G.J., Verkaik, E. 1999. De 10 meest gestelde vragen over koostofvastlegging in bos. Nederlands Bosbouwtijdschrift 71 (1): 2-6.

Verkaik, E., Nabuurs, G.J. 2000. Wood production potentials of Fenno-Scandinavian forests with nature conservation. Scandinavian Journal of Forest Research 15: 445-454.

Nabuurs, G.J., Dolman, A.J., Verkaik, E., Kuikman, P.J., van Diepen, C.A., Whitmore, A.P., Daamen, W.P., Oenema, O., Kabat, P., Mohren, G.M.J. 2000. Article 3.3. and 3.4. of the Kyoto Protocol - consequences for industrialised countries' commitment, the monitoring needs and possible side effects-. Environmental Science and Policy 3: 123-134.

Pussinen, A., Schelhaas, M.J., Verkaik, E., Heikkinen, E., Liski, J., Karjalainen, T., Päivinen, R., Nabuurs, G.J. 2001. Manual for the European Forest Information Scenario Model (EFISCEN 2.0). European Forest Institute, internal report No. 5 (available at www.efi.fi).

Nabuurs, G.J., Päivinen, R., Schelhaas, M.J., Pussinen, A., Verkaik, E., Lioubimow, A., Mohren, G.M.J. 2001. Long-term effects of nature oriented forest management in Europe. Journal of Forestry 99 (7): 28-33.

Verkaik, E., Jongkind, A.G., Berendse, F. Short-term and long-term effects of tannins on nitrogen mineralisation and litter decomposition in kauri (Agathis australis (D. Don) Lindl.) forests. Plant and Soil, in press.

111

Printed by Ponsen & Looijen, Wageningen.

The investigations were supported by the Research Council for Earth and Life Sciences (ALW) with financial aid from the Netherlands Organisation for Scientific Research (NWO).

112