University of Eastern Finland.

Luonnontieteiden ja Metsätieteiden tiedekunta.

Faculty of Science and Forestry.

THE EFFECT OF FOREST SOIL PRODUCTIVITY AND VEGETATION

TYPE ON DIPRIONID SAWFLIES COCOON PREDATION

Mar Ramos Sanz

MASTER´S THESIS OF SCIENCE IN AGRICULTURE AND FORESTRY

SPECIALITATION IN FOREST ECOLOGY AND SILVICULTURE

2

JOENSUU 2016

3

Ramos Sanz, Mar. 2016. The effect of Forest soil productivity and vegetation type on

Diprionid sawfly cocoon predation. University of Eastern Finland, Faculty of Science and

Forestry, School of Forest Sciences, Master thesis of Science in Agriculture and Forestry,

specialization in Forest Ecology and Silviculture, 52p.

ABSTRACT

The abundance of insect populations can change dramatically from generation to generation,

and large increases are commonly known as "outbreaks". Insect outbreaks can be extremely

destructive when the insect is considered as a crop or forest pest or it carries disease to

humans, farm animals, or wildlife. Due to the economic losses caused by pests, it is very

important to know and understand the processes (biotic or abiotic) that regulate insect

populations. In Finland, pine sawflies (Hymenoptera, Diprionidae), consisting mainly in two

species; Diprion pini (L.) and Neodiprion sertifer (Geoffroy) are one of the major defoliator

groups of Scots pine forests. Several studies have recently estimated future outbreaks of pine

sawflies, in the light of these future threats researchers considered that there is an increasing

need in study their main controllers.

In this research, I studied one of the most important agents controlling pine sawflies

populations, the cocoon predators, and how they are affected by the environment. Among

other hypothesis and theoretical background, I took into account the Exploitation Ecosystem

Hypothesis (EEH), which makes different predictions for predation depending on productivity

levels. Hence, I focused my study on the different vegetation types of Scots pine forests. I

used empirical models for the diprionid cocoon predation pressure, specifically generalized

linear mixed model (GLMM) techniques. The mean percentage of predation was predicted as

a function of forest vegetation type and season. The results of this research show that

predation pressure is highly related with the type and structure of the forest where the pine

sawflies predators live and predate. Rich forests (Mesic forests in this study) with higher

vegetation diversity and structures supported the highest levels of pine sawflies cocoon

predation.

Keywords: insect outbreaks, pine sawflies, predators, forest vegetation type and Exploitation

Ecosystem hypothesis (EEH).

4

FOREWORD

This thesis is the final project for a M.Sc. of Agriculture and Forestry specialization in Forest

Ecology and Silviculture, degree at the University of Eastern Finland. The thesis was

conducted in the laboratories of Metla building, which form part of the Finnish Natural

Resources Institute (Luke) and in Scots pine stands owned by Paihola Forest Common and

UPM, located near Joensuu (Finland). The advisory committee has two members: Dr. Seppo

Neuvonen (Luke) and Dr. Olli-Pekka Tikkanen from the University of Eastern Finland. I

would also like to express my gratitude to all the people from Metla building for their help

and technical support, and Antonio Rodríguez Olmo for his consultation and advice.

5

TABLE OF CONTENTS

1. INTRODUCTION………………………………………………..………….....…….…6

2. 1. Background and objectives………………………………………………………......6

1. 1. 1. Insect outbreaks and their importance in Forest ecosystems…………...……….....6

1. 1. 2. Forest insect outbreaks and main hypotheses of herbivore control.……......…..…..7

1. 1. 3. Pine sawflies and their control……………..…………….……………………..…..9

1. 2. Hypothesis and specific objectives……………………………………………….....12

2. MATERIAL AND METHODS…………………………………………………...…..13

2. 1. Insect cultures……………………………...……………………….……………...….13

2. 2. Study insect life cycles………………………………………………………..…...….15

2. 3. Study sites………………………………………………………………………......…16

2. 4. Experimental design………………………………………………………….……….17

2. 5. Pupae/Cocoon analysis………………………………….………………….……........19

2. 6. Statistical analysis…………………………………………………………………....21

2. 6. 1. Data analysis and modeling………………………………………………….….....21

3. RESULTS……………………………………………………………….………..……..24

3. 1. Total predation…………………………………………...………..………….............26

3. 2. Vertebrate predation………………………………………………………………......29

3. 3. Invertebrate predation....................................................................................................31

4. DISCUSSION...................................................................................................................33

5. CONCLUSIONS..............................................................................................................37

6. REFERENCES................................................................................................................38

7. APPENDIX......................................................................................................................47

6

1. INTRODUCTION

Variations in space and time of host plant quality can affect the effectiveness and behavior of

insect herbivores populations, ending in severe cases, to explosive increases in abundance

(outbreaks) of those insect species (Berryman 1999). Herbivores during these periodic

oscillations or outbreaks, can in return, influence their host resource availability (Krause &

Raffa 1996). Such interactions may be complicated by the presence and interactions with

natural enemies, which can in some cases mediate outcomes of plant–insect associations at

the individual and population levels (Power 1992, Krause & Raffa 1996). In this context, and

considering the ecological and economical importance of insect outbreaks, a better

understanding of how these interaction on tri-trophic systems, would be a key issue in future

forest pest management, through the development of more effective, and environmentally

benign natural integrated pest-management strategies (Krause & Raffa 1996).

1. Background and objectives

1. 1. 1. Insect outbreaks and their importance in forest ecosystems.

An outbreak can be defined as an explosive increase in the abundance of a particular species,

commonly called pests, which occur over a relative short period of time (Berryman 2003,

FAO 2010, Barbosa et al. 2012). In a broad sense, a pest is an insect that can damage or kill,

during some stage of their development; agricultural crops, forest stands, ornamental plants,

or native plants in situ, consume and/or damage harvested wood, cause illness or

unproductivity in agricultural animals (i.e: cattle), or be vector of human diseases (Berryman

1987, Wallner 1987, Liebhold et al. 1995, Liebhold & McCullough 2011).

Temporal and spatial oscillations in abundance, or “outbreaks”, are one of the most significant

characteristics of animal and insect population dynamics (Liebhold et al. 1995). These cycles

affect the ecosystems in which insect populations live and develop and are common in many

forest insect populations (Esper et al. 2007, Björkman et al. 2011). Impacts of forest insect

outbreaks include extensive defoliation and/or tree mortality concluding in many cases with

decreases in forest productivity and carbon storage, affecting forest structure, species

composition and other ecosystem patterns (Niemelä et al. 2001, Jepsen et al. 2008). Outbreaks

of insect herbivores causing a intermediate disturbances are known to act as important drivers

of forest dynamics (Hunter 2001, Kouki et al. 2001, Behmer 2009).

7

There are many more factors affecting these cycles, however measuring and evaluating their

influence is a difficult task (Coviella et al. 1999, Bale et al. 2002, Aukema et al. 2006,

Tscharntke et al. 2007). In the case of climate, the intricacy of those ecological interactions

are really difficult to quantify or predict especially changes in ecosystem processes (Mattson

& Haack 1987, Volney & Flemming 2000, Logan et al. 2003). The problem in factors such as

time and space is that detection and evaluation of those changes needs the measurement of

long time series, and wide geographical ranges which are not commonly available for most

natural systems (Ovaskainen & Hanski 2002, Hastings 2004, Koons et al. 2005, Yakamura et

al. 2006).

Taking into account not only damages caused by defoliation, but also impacts of forest insects

affecting nutrient cycle in the forest ecosystem, it can be considered that worldwide insect

pests affect around 35 million hectares of forests each year, with their consequent economic

losses (Allen et al. 2010, Alalouni et al 2013, Haynes et al. 2014). Economic losses caused by

defoliating pest insects can be substantial (~300 - 1000 EUR ha-1 depending on intensity of

needle loss and the length of outbreak period) (Latva-Käyrä 2011). Thus, enhancing the

understanding of the biological controllers of these insects, and their potential success in

controlling and regulating these herbivores population has become a significant topic in the

field of forest research.

1. 1. 2. Forest insect outbreaks and main hypotheses of herbivore control.

Along history, many insect species had caused severe losses related with their outbreaks, yet

scientists are still trying to understand the outbreak phenomena (Berryman & Turchin 2001,

Berryman 2003, Price et al. 2005). During decades, insect pests have provided to insect

ecology researchers with many challenges (Kendall et al. 1998, Liebhold & Tobin 2008,

Alalouni et al 2013).

Many hypotheses have been formulated in an effort to explain the causes of insect outbreaks

(Berryman 2003, Kessler et al. 2012, Stenberg 2015). Multitude of negative feedbacks from

lower (e.g. host plants, prey) and/or higher trophic levels (e.g. predators, diseases) are able to

produce insect outbreaks, if the feedback is impacting after a time lag (Berryman 1996).

However, in general, researchers agree that most oscillating cycles and outbreaks in forest

pests, are the result of a combination of trophic interactions that contribute delaying negative

8

feedbacks (Ginzburg & Taneyhill 1994, Hunter et al. 1997, Kendall et al. 1999, Hanski et al.

2001, Kessler & Baldwin 2002, White 2006).

Along the history, two completely opposite approaches have been proposed to explain the

population dynamics of herbivores (Moreau et al. 2006). Top-down approach point out that

herbivore populations primarily by the trophic level above (i.e. natural enemies) (Hairston et

al. 1960). This first approach was presented by Hairston et al. (1960) and although, they did

not pose the question in their article it has been translated as “why the world is green?” till

now (Hairston et al. 1960, Matson & Hunter 1992, Bond 2005).

In contrast to the top-down approach presented by Hairston et al. (1960), the bottom-up

approach (Mattson & Addy 1975) suggested that the trophic level below (i.e. the plant

resource) is the principal limitation of herbivore populations. Thus, bottom-up hypotheses,

consider that herbivores and other organisms are resource limited (e.g., low host quality,

presence of repellent or deterrent chemicals on their host plants etc.), even if their hosts plants

appear to be abundant (Janzen 1988, Kessler & Baldwin 2002). Bottom-up hypotheses

considered that herbivores do not have a regulating effect or any significant influence on the

productivity of their host plants (Denno et al. 2002, Price et al. 2005, Turkington 2009).

Another view, consisted in considering that both, bottom-up and top-down processes can act

together to influence the dynamics of herbivore populations (Hunter et al. 1997). This new

joint hypothesis has become largely accepted recently (Moreau et al. 2006). According to this

new concept, the instability on the herbivore population dynamic responses, could rely on

several attributes of the community (Hunter 2001). Including both, bottom-up stimulus such

as food quality of the host plant, and top-down stimulus such as the behaviour, abundance and

success of natural enemies (Coupe & Cahill 2003). However, there are still remaining

questions relating to the stability among these trophic forces, and how changes in the

ecosystem can affect this stability (Rosenheim 1998, Terborgh et al. 2001, Turkington 2009,

Kollberg et al. 2013).

Primary productivity has been postulated as one of the many factors modulating the relative

strengths of bottom-up and top-down forces on herbivore population dynamics (Coupe &

Cahill 2003). Related to this Oksanen et al. (1981) described the Exploitation Ecosystem

hypothesis (EEH) that converges with Hairstorn´s hypotheses with respect to productive areas

(forests and their successional stages, productive wetlands etc.). According to EEH, the

control of herbivores by predators declines in unproductive ecosystems (tundras, high alpine

9

areas, steppes and semi deserts among others) which can be then characterized by their

intense natural folivory (Oksanen et al. 1981, Oksanen & Oksanen 2000).

Thus, in low productivity environments plant biomass will be limited by nutrient availability

and herbivore populations effect on biomass regulation will be low (Oksanen et al. 1981). In

intermediate environments herbivore populations will be sustained by plants but not predator

populations and the system will be dominated by plant-herbivore interaction (Oksanen &

Oksanen 2000, Turkington 2009). Finally, in rich systems, plants can withstand both

populations of herbivores and predators, and in those environments the system will be

regulated by predator-prey interaction where predators will keep herbivore populations low

enough to have little impact on the plant biomass (Oksanen & Oksanen 2000).

1. 1. 3. Pine sawflies and their control.

Pine sawflies are among the most common defoliating insect species of pines forests all over

Europe (Larsson & Tenow 1984, Lyytikäinen 1994, Virtanen et al. 1996, Augustaitis 2007,

Dukes et al. 2008). Outbreaks are usually followed by long periods of low population

densities for a number of decades or longer, although the species are still present in the forest

ecosystem (Niemelä et al. 2001, De Somviele et al. 2004). Pine forests vary in their

susceptibility to sawflies invasions, and outbreaks are more common in pine forests on low

productivity soils (Niemelä et al. 2001, De Somviele 2004; 2007, Jepsen et al. 2008, Allen et

al. 2010, Björkman et al. 2011, Nevalainen et al. 2015).

Pine sawfly larvae of Diprion pini (L.) and Neodiprion sertifer (Geoffroy) (Hymenoptera:

Diprionidae), are the most common defoliating sawflies of Scots pine trees, Pinus sylvestris

L., in Finland (Juutinen & Varama 1986, Niemelä et al. 1991, Virtanen et al. 1996,

Lyytikäinen-Saarenmaa 1999). Outbreaks typically occur within 10-20 years intervals in

central Europe (Hanski 1987, Pschorn-Walcher 1987, Larsson et al. 2000, Anderbrant 2003,

Kurkela et al. 2005). According to Berryman (1987), the most common type of pine sawfly

outbreak is sustained and eruptive, which is defined as; an outbreak that spreads from local

epicentres to cover large areas and persists for several to many years (Larsson & Tenow 1984,

Berryman 1987, Virtanen et al. 1996).

Among other damages, large sawfly infestations can cause growth loss and mortality,

especially when followed by attacks from bark and wood-boring beetles (Coleoptera:

10

Buprestidae, Cerambycidae, Scolytidae,) (Kolomiets et al. 1979, Neuvonen & Niemelä 1991).

Pine sawflies species, previously considered as harmless species, are now causing severe

damage in Finnish forests (Lyytikäinen-Saarenmaa & Tomppo 2002, de Somviele et al.

2007), not only because of their rise and proliferation in affected areas, but also because of

their high damage potential (Kouki et al. 2001, Niemelä et al. 2001, Barre et al. 2002, Kantola

et al. 2011).

Pine sawflies spend part of their development as a pupae on/or nearby the forest soil, and this

life history trait exposes them to predation by ground-dwelling vertebrates and invertebrates

(Pschorn- Walcher 1987, Björkman & Gref 1993, Lyytikäinen 1994, Kytö et al. 1999,

Lyytikäinen-Saarenmaa 1999, Lee et al. 2001, Baldassari et al. 2003, Lindstedt et al. 2006,

Björkman et al. 2011). Despite environmental factors and host plant effects, predators (mainly

small mammals, insects, and birds etc.), pathogens and parasitoids are thought to be the more

important factors regulating and controlling pine sawflies populations (Hanski & Parviainen

1985, Heliövaara et al. 1991, Kouki et al. 1998, Larsson et al. 2000, Herz & Heitland 2005,

Stenberg 2015).

The main important research on diprionid cocoon predators was done by Hanski in the 90´s

decade. According to his work the control of a prey at low density requires the operation of

one or more directly density-dependent factors, for example predation (Hanski 1990b). In this

context one may be attempted to dismiss predation by shrews and other omnivorous mammals

as unimportant when no density dependence has been detected (Hanski 1990a).

However, considering the spatial element in outbreaks Hanski suggested that sawfly

populations should had been examined in the context of metapopulation scenario in which

small mammals might play a major role in the control of pine sawflies. This hypothesis is

commonly known as the “generalist predator theory” or more generally the “metapopulation

model of forest insects dynamics” (Hanski 1990a, Hanski & Hettonen 1996, Hanski et al.

2001).

Related with Hanski´s work Oksanen et al. (1981) proposed their Exploitation Ecosystem

hypothesis (EEH), in which they pose that different predictions on predation and pest control

could be obtained considering the dependence upon productivity levels on the ecosystem in

which they operate (Oksanen et al. 1981, Oksanen & Oksanen 2000).

11

Both hypothesis pointed that predation pressure and effectiveness may be significantly

affected by the habitat in which predators operate, the structure of the forest floor, the

availability of food resources, affecting consequently on their success and increasing the

probability risk of outbreaks in certain insect species (Oksanen et al 1981, Hanski

1987;1990a, Kollberg et al. 2013). I took into account these two hypotheses and their

theoretical background to make my own hypothesis and tried to answer the ecological

interactions studied in this thesis.

I studied the cocoon predation pressure and effectivity, taking into account not only the

presence and abundance of prey but also the quality of the environment in which the predators

and preys are located. In my knowledge, there are no previous studies where, the effect of all

cocoon predators (vertebrates and invertebrates), that potentially regulate pine sawflies

populations of both species, are considered, and the effect of the environment in their

predation success is quantified. In this study the vegetation type that describes the

productivity of the forest stands was hypothesized as a key factor in the success of the control

of pine sawfly populations by cocoon predation.

12

1. 2. Hypothesis and specific objectives.

The main hypothesis of this study is:

“High productivity forest soils (mesic vegetation type) are expected to have higher

predation pressure on pine sawfly cocoons than lower productivity forest soils (xeric and

sub-xeric vegetation types)”

Specific objectives:

1) Analyse the cocoon predation response considering all types of predators in three

different forest vegetation types.

2) Analyse the cocoon predation response of vertebrate predators in three different forest

vegetation types.

3) Analyse the cocoon predation response of invertebrate predators in three different

forest vegetation types.

4) Analyse the different predation responses of vertebrate and invertebrate predators.

13

2. MATERIAL AND METHODS

2. 1. Insect cultures

For conducting this experiment approximately 3000 larvae of Neodiprion sertifer (Geoffroy)

(Hymenoptera: Diprionidae), were reared during the summer of 2014. The larvae were from

different locations near Joensuu and Puumala (this last location was chosen because it

maintains a permanent population of the insects). The insect were disposed in groups of 20-30

larvae of same instar of development in plastic boxes with food ad libitum (Scots pine shoots

and cut branches) until they reached the pupae instar (cocoon) (see Figure 1; cf. Baldassari et

al. 2003). The rearing took place in the laboratories of Metla building (Joensuu) in a period of

two months, from 26 of May to 31 of July.

The plastic boxes for rearing were divided in three layers from bottom to top (see Figure 1):

1. The bottom layer consisted on Sphagnum and other mosses previously treated in order

to avoid soil predators and other biotic agents (Sphagnum and soil were boiled and

then dried).

2. The second layer was a wet filter paper with a threefold purpose; keeping a moist

environment inside the box, preventing contamination of the Sphagnum layer by the

larval faeces, and setting an easy to clean division between the larvae and the

Sphagnum layer, where larvae can move but no their faeces (some larvae prefer to

pupae inside the Sphagnum layer) and because the cleaning of faeces was easier.

3. The third layer contained the larvae and the plant material (Scots pine shoots and cut

branches) which was previously washed and checked to avoid the presence of

predators such as spiders, beetles or others.

4. Finally the top layer consisted of net cloth that allowed the entrance of air in the box

while preventing the escape of larvae.

Every second day during two months boxes were checked and faeces were removed from the

middle layer, the filter paper was changed, dead larvae were removed and fresh new plant

material was changed and added. The rearing boxes were maintained clean in order to avoid

contamination and larval mortality (Olofsson 1987, Heliövaara et al. 1991, Björkman & Gref

1993, Larsson et al. 2000, Giertych et al. 2007).

14

The boxes were placed in the laboratory at room temperature (18±2ºC) with a natural

photoperiod and their position in the room were changed randomly every day (Barre et al.

2002, Baldassari et al. 2003, Kollberg et al. 2013) (see Figure 1).

Figure 1. Neodiprion sertifer larvae (a) and pupae (b), a rearing box presenting the three first

layers and (c) and a rearing box with all the layers (d) (Source: Mar Ramos Sanz).

At the end of the rearing season (June-July 2014) approximately 2.600 pupae (cocoons) of N.

sertifer were collected. These cocoons were divided to conduct different experiments and

placed in a dark cold room (5ºC±1 ºC) to prevent the hatching of the adults (Baldassari et al.

2003, Kollberg et al. 2013). Considering the information obtained in previous studies, a high

probability of larval mortality was expected during the rearing period due to baculovirus and

fungi (Kolomiets et al. 1979, Olofsson 1989, Morimoto & Nakamura 1989, Heliövaara et al.

1991, Saikkonen & Neuvonen 1993, Lyytikäinen-Saarenmaa 1999).

15

Thus, larvae were carefully treated while rearing. Whenever one or few disease larvae were

detected they were discarded in order to avoid a quick spread of the disease. The final result

shows the importance of a careful rearing: at the end of the rearing period the mortality rate of

larvae due to fungi and baculovirus was only 8.3±2%.

2. 2. Study insects life cycles

I used the previously collected N. sertifer cocoons (1200 cocoons) for studying the predation

pressure of both species (see Figure 2.). I could not use cocoons of D. pini because I did not

had access to larvae of this species. However both species have similarities in their life cycles,

and morphologies of their developmental stages (specially their larvae and pupae). Because of

that, I took advantage of those similarities, and thereby obtain results about predation pressure

on both species.

N. sertifer lays eggs on Scots pine needles at the beginning of spring, their larvae fed on

groups in Scots pine needles during spring time and in early summer they begin to pupate, the

pupation period takes around one month and on later summer adults begin laying eggs which

overwinter in this state of development (see Figure 2.) (Kolomiets et al. 1979, Hanski 1987,

Lyytikäinen-Saarenmaa 1999, Pasquier-Barre et al. 2000).

On the other hand, D. pini, lays eggs during mid and late summer, their larvae fed on Scots

pine needles during summer till early autumn when it develops into pupae overwintering in

pupal stage (see Figure 2.) (Sharov 1993, Neuvonen & Niemelä 1991). Furthermore, D. pini

spend the main part of its cycle in the cocoon stage not only hibernating but sometimes even

diapausing up to several years (Herz & Heitland 2003, De Somviele et al. 2007). However

both cycles can be variable in time (1-1.5 months) in Finland due to environmental factors

(especially temperature).

There were a high possibility that many N. sertifer adults hatch approximately a month before

D. pini entered in pupal stage (see Figure 2.) (Dahlsten 1967, Hanski 1987, Pschorn-Walcher

1987, Neuvonen & Niemelä 1991, Sharov 1993, Kouki et al. 1998, Larsson et al. 2000). In

order to avoid this option, I decided to submit half of the pupae to a heating treatment (48

hours at 90ºC) and thus kill the individual inside pupal. Consequently, the insect inside the

cocoon was dead but the nutritional properties for predators were the same (Hanski 1990a).

16

Figure 2. Life cycle of Neodiprion sertifer and Diprion pini (E=Eggs; LA=Larvae;

PP=Prepupae; PU=Pupae, and AD=Adult) (source: Mar Ramos Sanz cf. Kolomiets et al.

1979, Sharov 1993, Lyytikäinen-Saarenmaa 1999).

2. 3. Study sites

Study sites were primarily even-aged Scots pine forests representing different vegetation

types on relatively mesic to dry soils. The majority of forests in the area were young or

middle-aged (approximately 40-50 years). To conduct the predation experiment two different

locations near Joensuu were used; Kruununkangas (62° 36' 55"N, 29° 55'E) and

Jaamankangas (62° 42´N, 29° 43'E) with three stands per locality dominated by Pinus

sylvestris of approximately the same age class per locality (see Figure 3 and Appendix 2).

In both locations, the stands of the following three soil/vegetation types were selected; Forest

growing on poor productivity soils (Xeric or Calluna type), forest growing on medium

productivity soils (Sub-xeric or Vaccinum type), and forest growing on high productivity soils

(Mesic or Myrtillus type) (see Figure 3, and Appendix 2). A complete description of the

stands is provided in the Appendix section (Appendix 2). I followed Cajander´s (1949),

vegetation units and forest vegetation types, to describe the study sites of this research

(Cajander 1949, Kuusipalo, 1983, Tonteri et al. 1990; 1990b, Tonteri 1994, EEA 2006).

Two rain gauges were established per locality and they were checked every week during the

time when the experiment was set, in order to consider the effect of precipitation in the

experiment (Kollberg et al. 2014). The summer and autumn temperatures and rain were

normal compared to previous years, and also the two rain gauges located close to both areas

follow this trend of precipitation (see Appendix 3, Figure 1).

17

Figure 3. Location of the study sites (Source: © Maanmittauslaitos (National Land Survey of

Finland 2010.).

2. 4. Experimental design

The aim of this experiment, consisted of measuring the cocoon predation pressure per site

while considering the effect of productivity and vegetation type of the chosen stands. In this

case, the chosen prey were two of the major defoliators of Scots pine forests, the sawflies

Neodiprion sertifer and Diprion pini (Hymenoptera, Diprionidae), however as I explained

above the cocoons used in the experiment were from N. sertifer previous rearing.

18

The cocoons/pupae collected during the rearing period were used as baits for all kind of

predators (vertebrate and invertebrate). These baits enclosed a 1 m long line of string in

which, every 20 cm a thread of around 15 cm with a glued pupae in the end was attached

(©Loctite super glue precision) (see Figure 4).

Simultaneously, a total of 50 cocoons were glued, reared and observed separately in order to

measure a possible side effect of the glue treatment. All of them developed normally. All the

pupae used in the experiment were randomly chosen (both sexes randomly mixed), glued to

thread ends, and attached to string lines. Every string had 5 threads attached with 5 cocoons

glued in their extremes (see Figure 4). Those strings were kept under laboratory conditions

(room temperature (18±2 ºC) with a natural photoperiod before the experiment was set in the

field.

As it was explained in the previous section, I decided to use half (600) of the obtained

cocoons for simulating N. sertifer with natural cocoons and half (600) for simulating D. pini

populations (see Figure 2 and Appendix 3). The first 600 live cocoons were set at the end of

June simulating the pupal stage of N. sertifer, and 600 heat treated cocoons at the end of

August simulating the pupal stage of D. pini (see Figure 2). The exposure of the cocoons were

4 weeks for each period. The length of the experiment was selected to permit the detection

and following activity of natural enemies (Niemelä et al. 1991, De Somviele et al. 2007).

Every cocoon was placed in the organic soil layer at approximately 3-5cm deep and cover by

the surrounding vegetation (Björkman & Gref 1993, Nageleisen & Bouget 2009).

In every Scots pine stand a 50m x 50m square grid was set and strings with attached cocoons

were placed in the four corners of the grid. Five strings (a total of 25 cocoons) were placed in

every corner at approximately 10 meters of the edge (see Figure 4 and Appendix 4). The

distance among strings was 20 cm (see Figure 4) and they covered an area of 1m2. Altogether,

a total of 100 pupae of N. sertifer were placed per stand and area, thus the final number of

pupae was 600 (200 per vegetation type) (see Appendix 4).

19

Figure 4. Diagram of the experiment placed in each Scots pine stand (left), example of a set

of 5 strings set in the field forming a 1m2 square (middle) and one string with its five threads

and pupae baits in detail (right).

After 4 weeks (at the end of June and in middle of September for N. sertifer and D. pini

experiments respectively), the cocoons were collected from the field and the remaining

cocoons of a single group of 5 strings with or without damages, were put separately in

labelled plastic bags, and placed in a dark cold room (5ºC±1ºC) for a later analysis

(Nageleisen & Bouget 2009).

2. 5. Pupae/Cocoon analysis

Once all the material was collected and placed in the laboratory, damages and signs of

predation in cocoons were classified in the following categories (see Figure 5):

- Intact pupae: Pupae without any visible damage, were considered non-predated.

- Disappeared pupae: Pupae disappeared from their thread were considered as

predated (Kouki et al. 1998, Denno et al. 2003).

- Regular hatching: The regular holes on one of the cocoon edges represented the

hatching of adult sawflies and they were considered as not predated (Kolomiets et

al. 1979, Nageleisen & Bouget 2009).

- Vertebrate predation: The damages included in this section were made by birds or

small mammals (voles and rodents). These damages involved big size irregular

holes in the external part of the cocoons, chewed cocoons, etc. (Kolomiets et al.

1979).

- Invertebrate predation: The main damages observed were produced by beetles.

These damages consisted in irregular or regular small size holes in the ventral

20

middle part of the cocoon (damages produced by the families Carabidae or

Elateridae) or irregular small size holes in one of the two edges of the cocoon

(mainly produced by family Staphylinidae) (Kolomiets et al. 1979, Nageleisen &

Bouget 2009).

- Parasitism: The damages produced by the hatching of parasitoids (small regular

holes or single regular holes in one of the cocoon extremes) were excluded of the

analysis and those cocoons were considered as non-predated. I took this decision

considering that I did not know the exact location of the larvae reared and due to the

fact that they could be already parasitized adding then noise to the results. (Herz &

Heitland 1999;2003).

Figure 5. Different types of damages and predation; a=Intact pupae, b= Hatched parasitoids

(considered as non-predated), c=Vertebrate damage (chewed cocoons), d= Adult sawfly hatch

hole, e=Small holes produced by hatched parasitoids, and f=Different types of invertebrate

damages in the middle-ventral side of the cocoons (mainly produced by beetles). (Source:

Mar Ramos Sanz).

21

2. 6. Statistical analysis

2. 6. 1. Data analysis and modelling

The collected general data is shown in Appendix 1. The number of total observations was 48

(each observation was obtained considering 4 square (with 25 pupae) x 3 vegetation type x 2

seasons), as a result the data had a hierarchical structure (Miina et al. 2009). The final data is

given in Table 1 (see also Appendix 1) in the following order; obs (number of observations),

forest vegetation type indexed as M (mesic), Sx (sub-xeric) and X (xeric), area

(J=Jaamankangas and K=Kruununkangas), site (this variable was included to avoid

overdispersion and included the effect of four 1m2 edges per stand as a random factor), season

(summer and autumn), total predation, vertebrate predation and invertebrate predation (see

Appendix 1).

The first general exploration of the data was done using descriptive statistical analysis. The

average and standard error of the percentage of the three response variables; intact pupae,

total predation, vertebrate predation and invertebrate predation were calculated for the three

vegetation types (Mesic, Sub-xeric and Xeric) and for the two seasons (summer and autumn).

I used these averages to represent graphically the distribution of the data of those three

response variables (divided in seasons and forest vegetation type). The mean percentage of

predation of both types of predation (vertebrate and invertebrate) was plotted together in order

to obtain a better representation of the results and to compare them. This first exploration

gave me an idea of the general responses of the treatments and allowed me to consider the

best way of analysis of the data (Quinn & Keough 2002).

Models were considered for total predation (Model 1), vertebrate predation (Model 2) and

invertebrate predation (Model 3). The percentage of predation values for these three

continuous variables were used as response variables (calculated by extracting the number of

cocoons consumed from the number of cocoons in each stand edge (25 cocoons, 1m2 square

subplots)). After that, I observed the distribution of those variables using their histograms (see

Appendix 5). None of these variables followed a normal distribution, instead of that they

followed a binomial distribution (Crawley 2005). Due to the binomial distribution of the data

set, I used statistical approaches that better match with the data obtained (Bolker et al. 2009),

instead of trying to fit the variables into classical statistical methods (using transformations to

obtain a normal distribution). Generalized linear mixed models (GLMMs) combine the

22

properties of two statistical frameworks that are widely used in ecology, linear mixed models

(which incorporate random effects) and generalized linear models (which handle non-normal

data by using link functions and exponential family e.g. normal, Poisson or binomial

distributions) (Venables & Ripley 2002, Bolker et al. 2009).

The response variables were modelled using generalized linear mixed models (GLMMs) with

a binomial response (logit-link function) because the data were collected in multiple levels of

grouping, and the predictors were both fixed (variables forest vegetation type and seasons),

and random (areas and sites) effects (McCulloch & Searle 2001). The multilevel hierarchy of

the data (edge, site, forest vegetation, area), and its subsequently correlated observations, was

taken into account by including random effects at different levels in the variance component

models, and by allowing the intercept to vary randomly across the levels (Miina et al. 2009).

Overdispersion of the data in the models was noted by adding a random term as being at the

bottom level (“pseudo” level), which in this case is the variable “site” (see Appendix 1, Table

1) (Venables & Ripley 2002, Miina et al. 2009, Rodríguez & Kouki 2015). The GLMMs were

estimated with maximum likelihood (Laplace approximation) using the cbind function of the

lme4 package for the R statistical programming language (R-3.2.2 version) (R development

Core Team 2013, Bates et al. 2015) to conduct binomial errors GLMMs.

For a better interpretation of the results and the possible interaction among predictors (forest

vegetation type and season). I checked their effects calculating the odds-ratio of the fixed

effects for the three models (using the formula: 1 − exp(𝑓𝑖𝑥𝑒𝑑(𝑚𝑜𝑑𝑒𝑙))). With the inclusion

of odds-ratio I could had a clearer interpretation of the fixed effects for all the levels for the

three models. Odds ratio were used to compare the differences among levels for both

predictors considering that they were calculated using the highest percentage as a baseline

(Miina 2009). Thus, odds-ratio showed the percentage of decrease between levels compared

with the highest level of predation detected. I also used the graphical representation to analyse

and interpret the results given by those odds ratio (Miina 2009, Bates et al. 2015, Rodríguez

& Kouki 2015).

23

I used the mean percentages of the observed data to plot two bar plots and show the

differences in percentage of predation (total, vertebrate and invertebrate) between forest

vegetation type and seasons. For doing this graphics I used MASS package in R for R 3.2.2

version (Crawley 2005, R development Core Team 2013). For representing graphically the

results of the models and the possible interaction among predictors (season and forest

vegetation type) I utilized sciplot package for R (R-3.2.2 version) (R development Core Team

2013, Morales 2015).

General model:

The general multi-level binomial model was utilized for the three models (total predation,

vertebrate predation and invertebrate predation) and was represented as follows (see Model

1):

𝑦𝑖𝑗𝑘~𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙(𝑛𝑖𝑗, 𝑝𝑖𝑗)

𝑙𝑛(𝑝𝑖𝑗|1 − 𝑝𝑖𝑗) = 𝑓(𝑥𝑖𝑗 , 𝛽) + 𝑢𝑖 + 𝑢𝑖𝑗 Model (1).

Where y is the observed predation in the edge (i.e. percentage of consumption calculated by

extracting the number of cocoons consumed from the number of cocoons in each stand edge

(25 cocoons, 1m2 square subplots)); 𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙(n, p) represents the binomial distribution with

its parameters𝑛 (binomial sample size, in this case 𝑛𝑖𝑗 are all equal to 25) and 𝑝 (proportion

of successes i.e. the number of cocoons consumed/predated); 𝑙𝑛(𝑝|1 − 𝑝) is a logit-link

function; and 𝑓(. ) is a linear function with arguments 𝑥𝑖𝑗 (i.e. fixed predictors) and β (i.e.

fixed parameters). Subscripts i and j refer to type of forest vegetation and season respectively.

ui, uij are random, normally distributed between-forest vegetation type and between seasons

with a mean of 0 and constant variances. Random terms at different hierarchical levels were

assumed to be uncorrelated (Bolker et al. 2009, Miina et al 2009).

24

3. RESULTS.

The calculated average and standard error for the general data is represented in Table 1. This

table shows the data separated in four subclasses; intact pupae (cocoons with no damage or

with signs of hatched adults from both parasitoids and sawflies), total predation (cocoons

predated by vertebrate and invertebrate), vertebrate predation (damages and signs of predation

produced by birds and small mammals) and invertebrate predation (damages and signs of

predation produced by insects mainly beetles). Thus, Table1 represents the mean percentage

of pupae and their predation per study site, (xeric, sub-xeric and mesic) and season (summer

and autumn).

General trend of data shows an increase in predation in mesic forest vegetation sites, the

maximum percentage (mean ± S.E) of total predation being in summer (51±7.6). Also the

maximum mean percentage of vertebrate predation appeared in mesic sites during summer

(35.5±11.5). However this trend cannot be applied to the mean percentage of invertebrate

predation that seems to be higher in autumn and mesic sites (23±4.6) (see Table 1).

Table 1. Proportions (%) of intact and total of predated pupae (mean ± S.E) for the two

seasons, summer (June) and autumn (September) and the three forest vegetation types (Xeric,

Sub-xeric and Mesic). Predation is given as total predation, vertebrate predation and

invertebrate predation.

Season site Intact Total

predation

Vertebrate

predation

Invertebrate

predation

Xeric 89.5±4.6 10.5±4.7 6±3.1 3.5±3

Summer Sub-xeric 82±4.5 18±4.5 13.5±4.8 4.5±1.2

Mesic 49±7.6 51±7.6 35.5±11.5 15.5±4.4

Xeric 81±1.4 19±1.4 4.5±1.59 14.5±1.8

Autumn Sub-xeric 68.5±5 31.5±4.5 15.5±6 16±4.47

Mesic 56±7.6 44±6.7 21±4.1 23±4.6

25

I analysed first the mean percentage results in summer, and the observed trend was consistent

with the main hypothesis of this study where mesic forest vegetation type had the highest

cocoon predation for all three types of predation (Total: 51±7.6; Vertebrate: 35.5±11.5;

Invertebrate: 15.5±4.4). In contrast, mean percentages of predation for Sub-xeric (Total:

18±4.5; Vertebrate: 13.5±4.8; Invertebrate: 4.5±1.2) and Xeric (Total: 10.5±4.7; Vertebrate:

6±3.1; Invertebrate: 3.5±3), forest vegetation types were much lower showing a lineal

decrease (see Figure 6).

When I explored the mean percentages for autumn period the results were less clear than for

the summer period, with almost no differences among mesic (Total: 44±6.7; Vertebrate:

21±4.1; Invertebrate: 23±4.6) and sub-xeric (Total: 31.5±4.5; Vertebrate: 15.5±6;

Invertebrate: 16±4.47) vegetation types, and showing differences between mesic and xeric

(Total: 19±1.4; Vertebrate: 4.5±1.59; Invertebrate: 14.5±1.8) but almost no differences

between sub-xeric and xeric (see Table 1 and Figure 6). One interesting result was that

invertebrate predation appeared higher in autumn than vertebrate predation for the three types

of vegetation (see Figure 6). In both cases (summer and autumn) vertebrates and invertebrates

presented a similar pattern with a higher predation in mesic forest vegetation type and lower

for sub-xeric and xeric (see Table 1 and Figure 6).

0

10

20

30

40

50

60

70

Mesic Sub-xeric Xeric

perc

enta

ge o

f pre

dation

Forest vegetation type

Summer

Total predation Vertebrate predation Invertebrate predation

26

Figure 6. The mean percentages (±S.E.) of total predation for the two seasons; summer (June-

July; above) and autumn (September; below).

3. 1. Total predation:

After fitting this first model for total predation, I firstly analysed the fixed factors (forest

vegetation type and season) and their interaction separately without analysing their levels.

This first approach was done using the estimated values of the model and analysing them

using a Chi-squared test. The results showed a significant effect of forest vegetation type (p-

value<0.0001) and an interaction among both fixed factors (p-value=0.0284) (see Table 2).

Table 2. General results of the Chi-squared analysis for estimated values for fixed factors

(Veg. type= Forest vegetation type and season) and their respective interactions, the df for

those values, the Sum of squares of the test, the Mean of squares and the corresponding p-

value.

Variable Df Sum of squares Mean squares p-value

Veg. type 2 36.74 18.37 <0.0001

Season 1 3.17 3.17 0.39

FxSeason 2 6.32 3.15 0.0284

I decided to use the season levels (summer and autumn) to contrast and obtain a better

interpretation of the results. In both cases, forest vegetation type appeared significantly

0

10

20

30

40

50

60

Mesic Sub-xeric Xeric

perc

enta

ge o

f pre

dation

Forest vegetation type

Autumn

Total predation Vertebrate predation Invertebrate predation

27

different, where the highest mean predation is presented in mesic forest vegetation level (the

reference level) (see Table 3). Sub-xeric forest vegetation type appeared significant different

(p-value<0.0001) compared with the other two types of forest vegetation, xeric and mesic (see

Table 3). Xeric vegetation type did not present significant differences compared with sub-

xeric vegetation type and mesic (p-value=0.081), as it is showed in Figure 7 and Table 3 (see

Table 3 and Figure 7).

The fixed factor season did not present significant differences among levels, where summer

appeared to have the highest percentage of total predation and was used as a reference level

(autumn p-value=0.40). However, season presented interaction with forest vegetation levels

sub-xeric (Sx:Season p-value=0.042) and xeric (X:Season p-value=0.028) (see Table 3).

Table 3. Statistical results for the first multilevel binomial model using mean percentage of

total predation as response variable with the estimated values for the fixed factors; Forest

vegetation type (M; Mesic, Sx; Sub-xeric, X; Xeric) and season (summer and autumn) and

their respective interactions (vegetation type and season), the standard error for those values,

the z-value of the test, the calculated odds-ratio and the corresponding p-value.

Variable Estimate Std

Error

z-value Odds

ratio

p-value

Forest

vegetation

type

Season

Interaction

M (reference level)

Sx

X

Summer (ref. level)

Autumn

M;Season (ref.level)

Sx:Season

X:Season

Total Predation

-0.57

-1.24

-0.27

-1.21

-1.40

0.41

0.42

0.35

0.56

0.63

-1.37

-2.92

-0.82

-2.03

-2.19

0.83

0.93

0.3

0.7

0.75

<0.0001

0.081

0.40

0.042

0.028

28

The highest predation was found in mesic forest vegetation types for both season summer and

autumn (see Figure 7). Compared to mesic forests vegetation type total predation was 83 %

lower in sub-xeric forests (odds-ratio=0.83) and 93 % lower in xeric forests (odds-ratio=0.93)

on average.

In autumn, compared to summer total predation was 30% higher (odds-ratio= 0.3) for all

forest vegetation levels on average (see Table 3 and Figure 7). Interaction showed a

significant decreased of total predation in sub-xeric forest vegetation sites 70% (odds-

ratio=0.7) and of 75% in xeric forest vegetation sites (odds-ratio=0.75) compared with mesic

forest vegetation type sites (see Table 3 and Figure 7).

Figure 7. The modelled mean percentage of total predation among seasons (summer and

autumn) and forest vegetation type (M=Mesic; Sx=Sub-xeric and X=xeric).

29

3. 2. Vertebrate predation:

In this second model for vertebrate predation, forest vegetation type predictor was

significantly different (p-value <0.0001) while season did not present significant differences

(p-value= 0.383) and there were not interaction between predictors (see Table 4).

Table 4. General results of the Chi-squared analysis for estimated values for fixed factors

(Veg. type= Forest vegetation type and season) and their respective interactions, the df for

those values, the Sum of squares of the test, the Mean of squares and the corresponding p-

value.

Variable Df Sum of squares Mean squares p-value

Veg. type 2 20.38 10.19 <0.0001

Season 1 0.46 0.47 0.383

FxSeason 2 0.79 0.4 0.602

Forest vegetation type presented significant differences among all levels. Thus, all the levels

showed different mean percentage values when they were compare to mesic forest vegetation

type (reference level) (sub-xeric p-value=0.042 and xeric p-value<0.0001) (see Table 5).

Table 5. Statistical results for the first multilevel binomial model using mean percentage of

vertebrate predation as response variable with the estimated values for the fixed factors;

Forest vegetation type(M; Mesic, Sx; Sub-xeric, X; Xeric) and season (Summer and Autumn),

the standard error for those values, the z-value of the test, the calculated odds-ratio and the

corresponding p-value.

Variable Estimate Std

Error

z-value Odds

ratio

p-value

Forest

vegetation

type

Season

M (reference level)

Sx

X

Summer (ref. level)

Autumn

Vertebrate predation

-0.91

-2.21

-0.26

0.45

0.5

0.56

-2.03

-4.43

-0.65

0.6

0.89

0.22

0.042

<0.0001

0.575

30

The highest vertebrate predation appeared in mesic forest vegetation sites for both seasons.

Compared to mesic forests there was a mean percentage decrease of 60% in sub-xeric (odds-

ratio=0.6) and 89% in xeric vegetation types (odds-ratio=0.89) (see Table 5 and Figure 8).

There were no significant differences among season levels (autumn p-value= 0.575) and no

interaction between levels of both predictors. Comparing summer (reference level) with

autumn the mean percentage of decrease was of 22% (odds-ratio=0.22) (see Table 5 and

Figure 8).

Figure 8. The modelled mean percentage of vertebrate predation among seasons (summer and

autumn) and forest vegetation type (M=Mesic; Sx=Sub-xeric and X=xeric).

31

3. 3. Invertebrate predation:

Finally, the third model showed significant differences for both predictor factors, forest

vegetation type and season with non-significant interaction among predictors (see Table 6).

Table 6. General results of the Chi-squared analysis for estimated values for fixed factors

(Veg. type= Forest vegetation type and season) and their respective interactions, the df for

those values, the Sum of squares of the test, the Mean of squares and the corresponding p-

value.

Variable Df Sum of squares Mean squares p-value

Veg. type 2 7.71 3.85 <0.0001

Season 1 16.2 16.19 <0.0001

FxSeason 2 4.31 2.15 0.35

Forest vegetation type was a significant predictor with significant differences between all

levels compared with mesic forest vegetation type (sub-xeric p-value<0.0001 and xeric p-

value<0.0001) (see Table 7). Season was also a significant predictor, showing significant

differences among summer and autumn, where autumn was the reference level (summer p-

value<0.0001) (see Table 7). For this third model differences were clearer than for the

previous two (see Figure 9).

Table 7. Statistical results for the first multilevel binomial model using mean percentage of

invertebrate predation as response variable with the estimated values for the fixed factors;

Forest (M; Mesic, Sx; Sub-xeric, X; Xeric) and season (Summer and Autumn), the standard

error for those values, the z-value of the test, the calculated odds-ratio and the corresponding

p-value.

Variable Estimate Std

Error

z-value Odds

ratio

p-value

Forest

vegetation

type

Season

M (reference

level)

Sx

X

Autumn (ref.

level)

Summer

Invertebrate predation

-1.04

-1.06

-1.28

0.36

0.36

-0.31

-2.85

-2.92

-4.02

0.64

0.65

0.72

<0.0001

<0.0001

<0.0001

32

The highest mean percentage of cocoons predated by invertebrates appeared in mesic

vegetation forest sites and during autumn (see Figure 9). When forest vegetation types levels

sub-xeric and xeric were compared with mesic, their odds-ratio calculated showed a decrease

in predation of 64% in sub-xeric forest vegetation (odds-ratio=0.64) and of 65% in xeric

forest vegetation type (odds-ratio=0.65) (see Table 7 and Figure 9). Cocoon predation by

invertebrates was significantly lower during summer (odds-ratio=0.72) for all forest

vegetation levels compared with autumn (see Table 7 and Figure 9).

Figure 9. The modelled mean percentage of invertebrate predation among seasons (summer

and autumn) and forest vegetation type (M=Mesic; Sx=Sub-xeric and X=xeric).

33

4. DISCUSSION.

This research considers three forest vegetation types ranging from poor sites (xeric or Calluna

type) to medium (sub-xeric or Vaccinum type) and rich sites (Mesic or (Myrtillus type) and

two main predator guilds (vertebrate and invertebrate). The predation pressure was considered

by summing both kind of predator types and separating them to observe which one was more

efficient and in which environment their effect was higher. The three applied models gave

similar results, all indicating that there were a significant difference in predation among the

three forest vegetation types especially between the extreme; the richer (Mesic) and the poorer

(Xeric).

Results obtained from this study strongly support the posed hypothesis and confirm the high

potential of cocoon predation as control agent of pine herbivores. This hypothesis is based in

the work of “generalist predator theory” of Hanski and the Ecosystem Exploitation hypothesis

(EEH) of Oksanen, and as expected forest on nutrient rich sites were populated by larger

populations of small mammals and other kind of predators and due to that predation pressure

was higher in those sites than in poor forest sites (Oksanen et al. 1981, Hanski 1990a).

However, this potential control can be threatened by the fact that there are no many natural

Scots pine forests growing in mesic forest vegetation types (Kuusipalo 1983, Tonteri 1994,

Fuller & Quine 2015).

The highest predation pressure was caused by vertebrate predators. However, since direct

observations of the predators affecting the cocoon population were not measured, I can only

speculate on the reasons behind these results. Previous studies pointed that small mammals

are the main vertebrate predators of this system, more specifically shrews and voles (Herz &

Heitland 2005). Shrews (Sorex spp.) and voles (Myodes spp. and Microtus spp.) are known to

find their food in places where it would be risky to stay for long periods, in example the open

forest floor (Hanski 1990a). Due to that, the food item is picked and transported to a spot

where the small mammal is safe from its predators and competitors (Hanski et al. 2001,

Sullivan et al. 2004). In this sense, when the gathering benefits fall below the energetic costs

and the risks of being killed are high, small mammals should stop foraging (Functional

response type 3) (Terborgh et al. 2001, Verdolin 2006). Opened and chewed cocoons were

mainly consumed by small mammals according to the specific feeding signs (Kolomiets et al.

1979, Raymond et al. 2002).

34

Both types of predation were observed in this study (cocoon disappeared or with feeding

signs). Due to the higher vertebrate predation pressure in mesic sites it can be considered that

these particular ecosystems provided small mammals enough shelter and food to developed

their activities reducing the risks of starvation and been predated, than in sub-xeric and xeric

sites (Hanski & Parviainen 1985, Hanski 1990a; 1990b, Larsson et al. 2001, Korpimäki et al.

2005), the number and quality of food captures have several important functions in the

ecology of shrews and other omnivorous small mammals (Hanski 1990b) promoting a

numerical response to a prey type that is otherwise effectively seasonal (Hanski 1990a).

Another important vertebrate predator of insect pests are birds (Morrison et al. 1990, Barbaro

& Battisti 2011). Kouki et al. (1998) pointed out in their discussion that birds could play a

very significant role in the functioning of forest ecosystems, and especially their importance

as insect defoliator regulators (Kouki et al. 1998). Birds prey on all sawfly stages, including

the egg, larval, cocoon and adult (Morrison et al. 1990, Barbaro & Battisti 2011). However

the actual effect of avian predation is not clear. In some habitats, most avian species highly

feed on sawflies but in other cases such predation has been viewed as insignificant (Morrison

et al. 1990). Avian predation may be particularly important at low prey population densities,

birds may also affect sawflies populations indirectly through the dispersal of pathogens

(Hanski & Parviainen 1985, Kouki et al. 1998, Barbaro & Battisti 2011, Kollberg et al. 2014).

Due to their likely possible relevance in this system I think birds should be considered in

future studies.

Despite that the main predation effect in this study was produced by vertebrates, invertebrate

predation was also significantly higher during autumn especially in mesic stands. According

with previous studies, invertebrate predation is mainly caused by beetles (Carabids and

Staphylinids among others) (Codella & Raffa 1993, Tanhuanpää et al. 1999 Raymond et al.

2002). Beetle predation leaves considerable fragments of the cocoon and make especial marks

such as irregular holes, whereas mammals remove the entire pupa from the cocoon or makes

different marks (Elkinton et al. 2004). In this research invertebrate predation was lower than

vertebrate predation but differed significantly among the three vegetation types especially for

35

the richer forest vegetation type (mesic), which had the highest invertebrate predation

pressure.

Although this experiment was not designed to measure the competition among predator

guilds, the general results gave a pattern in which both predators seem to have an association

ranging from no relationship to a negative relationship, having vertebrate predators a negative

impact over the potential cocoon predation of invertebrates (Larsson et al. 2000, Kollberg et

al. 2014). This association also could be explained as an inverse relationship between

vertebrate (mainly small mammals) and invertebrate predation (Tanhuanpää et al. 1999,

Raymond et al 2002, Elkinton et al. 2004). In that sense, small mammals would be more

effective in finding and foraging cocoons, and as a consequence invertebrate predators would

act attacking the remaining cocoons left by vertebrates, hence invertebrate predation probably

will be relatively higher when densities of small mammals are low (Lee et al. 2001, Finke &

Denno 2002; Hastings et al. 2002). Despite I cannot make a general statement, it is possible

that competition play an important role in this system and the relationship among both types

of predator guilds should be consider in future research.

In this study mortality caused by pupal parasitoids was not measured. The main reason was

that the larvae utilized in the experiment were originally sampled from different areas, outside

the study area. Because it was not controlled whether the larvae were previously parasitized or

they were parasitized as pupal stage during the experiment finally, it was decided not to

include parasitism in this study, and the pupae with marks of parasitism were counted as non-

predated. However based on the information found from previous studies, it is probable that

the influence of parasitism in the overall control of pine sawflies cocoons is not higher enough

compared with the control made by vertebrate predators (Hanski 1987, Herz & Heitland 1999;

2003; 2005, DeSomviele et al. 2007).

36

This study results showed a strong response in predation pressure especially in rich forests

(Mesic stands). Considering that both kind of predators (vertebrate and invertebrate) are

generalist, and usually do not present a strong numerical response to specific preys, but their

population densities are reasonably influenced by the abundance of all prey in the habitat

(Berryman 1994), this finding support the general knowledge of their ability to control low-

density insect populations (Hanski & Hettonen 1996). If this functional response can be

extend in time, those generalist predators have the potential of control and regulate this low-

density pine sawflies populations (Hanski 1990a, Kollberg et al. 2014).

37

5. CONCLUSIONS

Due to the potential economical, ecological and environmental damage that pine sawflies can

produce, during the last decades lot of research has been done in order to explore the specific

system that includes, pine sawflies and specially their predators and natural controllers. The

results of this master thesis study supports previous research suggesting, that apart from the

connections to weather, sawfly outbreaks often occur in forests growing on nutrient-poor

soils. This work demonstrates that predators and hence the pupal predation pressure is

significantly influenced by site productivity. The main explanation for this results is that food

resources are generally more abundant in rich forest habitats and this has a direct positive

influence on predator’s populations for several reasons, included that a rich forest usually

contains more sheltering structures which are used to escape from top predators. The most

important result of this research was the significant effect of forest vegetation type on both

kind of predators (vertebrate and invertebrate). As a result of this effect predation was higher

in mesic or richer forest vegetation types, than in sub-xeric or xeric poorer forest vegetation

types. Despite that the most important and effective type of predation was exerted by

vertebrate predators, it is relevant to notice that this is the first study in which both types of

predators (vertebrate and invertebrate) were included. Finally, it is important to consider that

this system is more intricate than it seems and whenever more variables and actors are

included (for example. in this study the weather, competition among predators and top

predators were not included) the more complex the results will be. Future research will be

focus on the effect of new factors (weather, competition, etc.) affecting the effectiveness of

the total predation.

38

6. REFERENCES

Alalouni, U., Schädler, M., and Brandl, R. 2013. Natural enemies and environmental factors

affecting the population dynamics of the gypsy moth. Journal of Applied Entomology. 137:721–738.

Allen, C. D., Macalady, A. K., Chenchouni, H., Bachelet, D., McDowell, N., Vennetier, M.,

Kitzberger, T., Rigling, A., Breshears, D. D. Hogg, E. H. T., Gonzalez, P., Fensham, R.,

Zhang, Z., Castro, J., Demidova, N., Lim, J-H., Allard, G., Running, S. W., Semerci, A., and

Cobb, S. 2010. A global overview of drought and heat-induced tree mortality reveals

emerging climate change risks for forests. Forest Ecology and Management. 259: 660–684.

Anderbrant, O. 2003. Disruption of pheromone communication in the European pine sawfly,

Neodiprion sertifer, at various heights. Entomologia Experimentalis et Applicata 107: 243–

246.

Augustaitis, A. 2007. Pine Sawfly (Diprion pini L.) – Related Changes in Scots Pine Crown

Defoliation and Possibilities of Recovery. Polish Journal of Environmental Studies, 16 (3):

363-369.

Aukema, B. H., Carroll, A. L., Zhu, J., Raffa, K. F., Sickley, T. A. and Taylor, S. W. 2006.

Landscape level analysis of mountain pine beetle in British Columbia, Canada:

spatiotemporal developments and spatial synchrony within the present outbreak. Ecography

29: 427441.

Baldassari, N., Martini, A., and Baronio, P. 2003. A technique for continuous rearing of

Neodiprion sertifer (Geoffroy) (Hym., Diprionidae) in the laboratory. Journal of Applied

Entomology, 127:103-106.

Bale, J. S., Masters, G. J., Hodkinson, I. D., Awmack, C., Bezemer, T. M., Brown, V. K.,

Butterfield, J., Buse, A., Coulson, J. C., Farrar, J., Good, J. E. G., Harrington, R., Hartley, S.,

Jones, T. H., Lindroth, R. L., Press, M. C., Symrnioudis, I., Watt, A. D., and Whittaker, J. B.

2002. Herbivory in global climate change research: direct effects of rising temperature on

insect herbivores. Global Change Biology. 8:1-16.

Barbaro, L., & Battisti, A. 2011. Birds as predators of the pine processionary moth

(Lepidoptera: Notodontidae). Biological Control. 56: 107–114.

Barbosa, P., Letorneau, D., and Agrawal, A. (eds). 2012. Insect Outbreaks Revisited. Willey

& Blackwell Ltd, Southern Gate, Chichester, West Sussex, UK .480p

Barre, F., Milsant, F., Palasse, C., Prigent, V., Goussard, F., and Géri, C. 2002. Preference

and performance of the sawfly Diprion pini on host and non-host plants of the genus Pinus.

Entomologia Experimentalis et Applicata 102(3): 229–237.

Bates, D., Maechler, M., Bolker, B., Walker, S., 2014. lme4: linear mixed-effect models using

eigen and S4 classes. R package, version 1.1-7. Available from

<https://github.com/lme4/lme4/> and <http://lme4.r-forge.r-project.org/>.

Behmer, S. T., & Joern, A. 2012.Insect herbivore viewed through a physiological framework:

Insights from Orthoptera. In: Insect Outbreaks Revisited (Barbosa, P., Letorneau, D. K., and

Agrawal A. A. Eds). Willey & Blackwell Ltd, Southern Gate, Chichester, West Sussex, UK.

4-23pp.

Berryman, A. A. 1987. The theory and classification of outbreaks. In: Barbosa, P., & Schultz

J. C. (eds). Insect Outbreaks. Academic Press, New York. 3-27p.

39

Berryman, A. A. 1999. Principle of Population Dynamics. Stanley Thornes Ltd., Cheltenham,

UK.

Berryman, A. A. 2003. On principles, laws and theory in population ecology. Oikos

103(3):695-701.

Berryman, A. A., & Turchin, P. 2001. Identifying the density-dependence structure

underlying ecological time series. Oikos. 92:265-270.

Bolker, B. M., Brooks, M. E., Clark, C. J., Geange, S. W., Poulsen, J. R., Stevens, M. H. H.,

and White, J-S. S. 2009. Generalized linear mixed models: a practical guide for ecology and

evolution. Trends in Ecology and Evolution. 24(3):127-135.

Bond, W. J. 2005. Large parts of the world are brown or black: A different view on the

‘Green World’ hypothesis. Journal of Vegetation Science, 16(3):261-266.

Björkman, C., & Larsson, S. 1991. Pine sawfly defense and variation in host plant resin acids:

a trade-off with growth. Ecological Entomology. 16: 283–289.

Björkman, C., & Gref, R. 1993. Survival of pine sawflies in cocoon stage in relation to resin

acid content of larval food. Journal of Chemical Ecology. 19(12):2881-2890.

Björkman, C., Bylund, H., Klapwijk, M. J., Kollberg, I., and Schroeder, M. 2011. Insect Pests

in Future Forests: More Severe Problems? Forests. 2:474-485.

Cajander, A. K. 1949. Forest types and their significance. Soumalaisen Krijalisuuden Seuran

Kirjapainen oy. Helsinki, 71pp.

Codella, S. G., & Raffa, K. F. 1993. Defense strategies of folivorous sawflies. In: Wagner, M.

& Raffa, K. F. (eds). Sawfly life history adaptations to woody plants. Academic Press, Inc.,

UK. 261-294.

Coupe, M. D., & Cahill, JR, J. F. 2003. Effects of insects on primary production in temperate

herbaceous communities: a meta-analysis. Ecological Entomology. 28:511-521.

Coviella, C. E., & Trumble, J. T. 1999. Effect of elevated atmospheric carbon dioxide on

insect-plant interactions. Conservation Biology. 13(4): 700-712.

Crawley, M. J. 2005. Statistics: An introduction using R. John Wiley & Sons Ltd, London,

UK. 333pp.

Dahlsten, D. L. 1967. Preliminary Life Tables for Pine Sawflies in the Neodiprion Fulviceps

Complex (Hymenoptera:Diprionidae), Ecology.48(2):275-289.

Daubenmire, R. 1976. The use of vegetation in assessing the productivity of forest lands. The

botanical review. 42(2): 115-143.

Denno, R. F., Gratton, G., Peterson, M. A., Langellotto, G. A., Finke, D. L., and Huberty, A.

F. 2002. Bottom-up Forces Mediate Natural-Enemy Impact in a Phytophagous Insect

Community. Ecology. 83 (5):1443-1458.

Denno, R. F., Gratton, C., Débel, H., and Finke, D. L. 2003. Predation Risk Affects Relative

Strength of Top-Down and Bottom-Up Impacts on Insect Herbivores. Ecology, 84(4):1032–

1044.

De Somviele, B., Lyytikäinen-Saarenmaa, P., and Niemelä, P. 2004. Sawfly (Hym.

Diprionidae) outbreaks on Scots pine: effect of stand structure, site quality and relative tree

position on defoliation intensity. Forest Ecology and Management 194: 305–317.

40

De Somviele, B., Lyytikäinen-Saarenmaa, P., and Niemelä, P. 2007. Stand edge effects on

distribution and condition of Diprionid Sawflies. Agricultural and Forest Entomology, 9:17-

30.

Dukes, J. S., Pontuis, J. Orwing, D., Garnas, J. R., Rodgers, V. L., Brazee, N., Cooke, B.,

Theoharides, K. A., Stange, E. E., Harrington, R., Ehrenfeld, J., Gurevitch, J., Lerdau, M.,

Stinson, K., Wick, R., and Ayers, M. 2008. Responses of insect pests, pathogens and invasive

plant species to climate change in the forests of the northeastern North America: What can

we predict? Canadian Journal of Forest Research. 39:231-248.

Elkinton, J. S., Liebhold, A. M., and Muzika, R-M. 2004. Effects of alternative prey on

predation by small mammals on gypsy moth pupae. Population Ecology. 46:171–178.

Esper, J., Büntgen, U., Frank, D. C., Nievergelt, D., and Liebhold, A. 2007. 1200 years of

regular outbreaks in alpine insects. Proceedings of the Royal Society B. 274:671-679.

European Environmental Agency. 2006. European forest types categories and types for

sustainable forest management reporting and policy. Technical report No 9/2006, European

Environment Agency, Copenhagen, Denmark.114p.

Finke, D. L., & Denno, R. F. 2002. Intraguild predation dismissed in complex-structured

vegetation: Implication for prey suppression. Ecology. 83(3): 642-652.

Food and Agriculture Organization of the United Nations. 2010. Global forest resources

assessment 2010: Main report. Food and Agriculture Organization of the United Nations,

Rome.

Fuller, L., & Quine, C. P. 2015. Resilience and tree health: a basis for implementation in

sustainable forest management. Forestry. 89:7–19.

Giertych, M. J., Karolewski, P., Grzebyta, J., and Oleksyn, J. 2007. Feeding behavior and

performance of Neodiprion sertifer larvae reared on Pinus sylvestris needles. Forest Ecology

and Management 242:700–707.

Ginzburg, L. R., & Taneyhill, D. E. 1994. Population Cycles of Forest Lepidoptera: A

Maternal Effect Hypothesis. Journal of Animal Ecology. 63: 79-92.

Hairston, N. G., Smith, F. E., and Slobodkin, L. B. 1960. Community structure, Population

Control and Competition. The American Naturalist. 94(879):421-425.

Hanski, I. 1987. Pine sawfly population dynamics: patterns, processes, problems. Oikos,

50:327-335.

Hanski, I. 1990a. Small mammal predation snd the population dynamics of Neodiprion

sertifer In: Watt, A.D., Leather, S. R., Hunter, M. D., and Kidd, N. A. (eds.). Population

dynamics of Forest insects. Intercept, Andover, Hampshire. 253-264.

Hanski, I. 1990b. Density dependence, regulation and variability in animal populations.

Philosophical Transactions of the Royal Society of London, B 330, 141-150.

Hanski, I., & Parviainen, P. 1985. Cocoon predation by small mammals, and pine sawfly

population dynamics. Oikos. 45: 125-136.

Hanski, I., & Hettonen, H. 1996. Predation on competing rodent species: a simple explanation

of complex patterns. Journal of Animal Ecology. 65: 220-232.

Hanski, I., Henttonen, H., Korpimäki, E., Oksanen, L., and Turchin, P. 2001. Small-Rodent

dynamics and predation. Ecology. 82(6):1505–1520.

41

Hassell, M. P. 2000. Host-parasitoid population dynamics. Journal of Animal Ecology.

69(4):543-566.

Hastings, A. 2004. Transients: the key to long-term ecological understanding? Trends in

Ecology and Evolution. 19: 39–45.

Hastings, F. L., Hain, F. P., Smith, H. R., Cook, S. P., & Monahan, J. F. (2002). Predation of

gypsy moth (Lepidoptera: Lymantriidae) pupae in three ecosystems along the southern edge

of infestation. Environmental entomology. 31(4):668-675.

Haynes, K. J., Allstadt, A. J., and Klimetzek, D. 2014. Forest defoliator outbreaks under

climate change: effects on the frequency and severity of outbreaks of five pine insect pests.

Global Change Biology.20: 2004–2018.

Heliövaara, K., Väisänen, R., and Varama, M. 1991. Larval mortality of pine sawflies

(Hymenoptera:Diprionidae) in relation to pollution level: A field experiment. Entomophaga.

36(2): 315-321.

Herz, A., & Heitland, W. 1999. Larval parasitism of a forest pest, the common pine sawfly,

Diprion pini (L.) (Hymenoptera. Diprionidae), during an endemic density phase. Journal of

applied Entomology. 123:129-137.

Herz, A., & Heitland, W. 2003. Impact of cocoon predation and parasitism on endemic

populations of the common pine sawfly Diprion pini (L.) (Hymenoptera: Diprionidae in

different forest types. Agricultural and Forest Entomology. 5:35-41.

Herz, A., & Heitland, W. 2005. Species diversity and niche separation of cocoon parasitoids

in different forest types with endemic populations of their host, the Common Pine Sawfly

Diprion pini (Hymenoptera: Diprionidae). European Journal of Entomology. 102: 217–224.

Hunter, M. D., Varley, G. C., and Gradwell, G. R. 1997. Estimating the relative roles of top-

down and bottom-up forces on insect herbivore populations: a classic study revisited.

Proceedings of the National Academy of Science. 94:9176–9181.

Hunter, M. D. 2001. Insect population dynamics meets ecosystem ecology: effects of

herbivory on soil nutrient dynamics. Agricultural and Forest Entomology.3: 77-84.

Janzen, D. H. 1988. On the Broadening of Insect-Plant Research. Ecology 69(4):905.

Jepsen, J. U., Hagen, S. B., Ims, R. A., and Yoccoz, N. G. 2008. Climate change and

outbreaks of the geometrids Operophtera brumata and Epirrita autumnata in subarctic birch

forest: evidence of a recent outbreak range expansion. Journal of Animal Ecology.77: 257–

264.

Juutinen, P. & Varama, M. 1986. Occurrence of the European pine saw fly (Neodiprion

sertifer) in Finland during 1966-83. Folia Forestalia 662: 1-39.

Kantola, T., Lyytikäinen-Saarenmaa, P., Vastaranta, M., Kankare, V., Yu, X., Holopainen,

M., Talvitie, M., Solberg, S., Puolakka, P., and Hyyppä, J. 2011. Using high density ALS data

in plot level estimation of the defoliation by the common pine sawfly. SilviLaser. 16-20.

Kendall, B. E., Prendergast, J., and Bjørnstad, O. N. 1998. The macroecology of population

dynamics: taxonomic and biogeographic patterns in population cycles. Ecology Letters. 1:

160–164.

Kendall, B. E., Briggs, C. J., Murdoch, W. W., Turchin, P., Ellner, S. P., McCauley, E.,

Nisbet, R. M., and Wood, S. N. 1999. Why do populations cycle? A synthesis of statistical

and mechanistic modeling approaches. Ecology 80: 1789–1805.

42

Kessler, A., & Baldwin, I. T. 2002. Plant responses to insect herbivory: The emerging

molecular analysis. Annual Review of Plant Biology. 53:299–328.

Kessler, A., Poveda, K., and Poelman, E. H. 2012. Plant-Induce responses and herbivore

population dynamics. In: Insect Outbreaks Revisited (Barbosa, P., Letorneau, D. K., and

Agrawal A. A. Eds). Willey & Blackwell Ltd, Southern Gate, Chichester, West Sussex, UK.

90-108pp.

Kollberg, I., Bylund, H., Schmidt, A., Gershenzon, J., and Björkman, C. 2013. Multiple

effects of temperature, photoperiod and food quality on the performance of a pine sawfly.

Ecological Entomology.38: 201–208.

Kollberg, I., Bylund, H., and Huitu, O. 2014. Regulation of forest defoliating insects through

small mammal predation: reconsidering the mechanisms. Oecologia. 176:975–983.

Kolomiets, N. G., Stadnitskii, G. V., and Vorontsov, A. 1979. The European pine sawfly:

Distribution, biology, economic importance, natural enemies and control. New Delhi:

American Publishing Co. 138p.

Koons, D. N., Grand, J. B., Zinner, B., and Rockwell, R. F. 2005. Transient population

dynamics: Relations to life history and initial population state. Ecological Modelling. 185:

283–297.

Korpimäki, E., Oksanen, L., Oksanen, T., Klemola, T., Norrdahl, K., and Banks, P. B. 2005.

Vole cycles and predation in temperate and boreal zones of Europe. Journal of Animal

Ecology. 74:1150-1159.

Kouki, J., Lyytikäinen-Saarenmaa, P., Henttonen, H., and Niemeläa, P. 1998. Coccon

predation on diprionid sawflies: the effect of forest fertility. Oecologia, 116:482–488.

Kouki, J., Löfman, S., Martikainen, P., Rouvinen, S., and Uotila, A. 2001. Forest

Fragmentation in Fennoscandia: Linking Habitat Requirements of Wood-associated

Threatened Species to Landscape and Habitat Changes, Scandinavian Journal of Forest

Research, 16(3): 27-37.

Krause, S. C., & Raffa, K. 1996. Defoliation tolerance affects the spatial and temporal

distributions of larch sawfly and natural enemy populations. Ecological Entomology. 21:259-

269.

Kurkela, T., Aalto, T., Varama, M., and Jalkanen, R. 2005. Defoliation by the Common Pine

Sawfly (Diprion pini) and Subsequent Growth Reduction in Scots Pine: A Retrospective

Approach. Silva Fennica, 39(4): 467-480.

Kuusipalo, J. 1983. Distribution of vegetation on mesic forest sites in relation to some

characteristics of the tree stand and soil fertility. Silva Fennica. 17(4): 403–418.

Kytö, M., Niemelä, P., Annila, E., and Varama, M. 1999. Effects of forest fertilization on the

radial growth and resin exudation of insect-defoliated Scots pines. Journal of Applied

Ecology. 36:763-769.

Larsson, S., & Tenow, O. 1984. Areal distribution of a Neodiprion sertifer (Hym.,

Diprionidae) outbreak on Scots pine as related to stand condition. Holartic Ecology. 7:81-90.

Larsson, S., Ekbom, B., and Björkman, C. 2000. Influence of plant quality on pine sawfly

population dynamics. Oikos. 89: 440–450.

Latva-Käyrä, P. 2011. Estimation of Scots pine defoliation by the common pine sawfly

(Diprion pini L.) using multi-temporal radar data. Metsävaratieteen ja –teknologian pro

gradu- tutkielma maatalous- ja metsätieteiden maisterin tutkielmaa varten. Helsinki, 62p.

43

Lee, J. C., Menalled, F. D., and Landis, D. A. 2001. Refuge habitats modify impact of

insecticide disturbance on carabid beetle communities. Journal of Applied Ecology. 38: 472–

483.

Liebhold. A. M., Macdonald. W. L., Bergdahl, D., and Mastro, V. C. 1995. Invasion by

Exotic Forest Pests: A Threat to Forest Ecosystems. Forest Science Monographs 30-49.

Liebhold, A. M., & Tobin, P. C. 2008. Population Ecology of Insect Invasions and Their

Management. Annual Review of Entomology.53: 387-408.

Liebhold, A.M. & D.G. McCullough. 2011. Forest Insects. In: Encyclopedia of Biological

Invasions (Simberloff, D. and M. Rejmanek, eds.) University of California Press, Berkeley

238-241.

Lindstedt, C., Mappes, J., Päivinen, J., and Varama, M. 2006. Effects of group size and pine

defence chemicals on Diprionid sawfly survival against ant predation. Oecologia. 150:519–

526

Logan, J. A., Régnière, J., and Powell, J. A. 2003. Assessing the impacts of global warming

on forest pest dynamics. Frontiers in Ecology and the Environment 1: 130–137.

Lyytikäinen, P. 1994. Effects of natural and artificial defoliations on sawfly performance and

foliar chemistry of Scots pine saplings. Annals of Zoologica Fennici 31:307-318.

Lyytikäinen-Saarenmaa, P. 1999. The Responses of Scots Pine, Pinus sylvestris, to Natural

and Artificial Defoliation Stress. Ecological Applications. 9(2):469-474.

Lyytikäinen-Saarenmaa, P., & Tomppo E. 2002. Impact of sawfly defoliation on growth of

Scots pine Pinus sylvestris (Pinaceae) and associated economic losses. Bulletin of

Entomological Research. 92:137-140.

Matson, P. A., & Hunter, M. D.1992. Special Feature: The Relative Contributions to Top-

Down and Bottom-Up Forces in Population and Community Ecology. Ecology 73:723.

Mattson, W. J. & Addy, N. D. 1975. Phytophagous insects as regulators of forest primary

production. Science.190:515–522.

Mattson, W. J., & Haack, R. A. 1987. The role of drought in outbreaks of plant-eating insects.

BioScience. 37(2): 110-118.

McCulloch, C. E., & Searle, S. R. 2001. Generalized, linear, and mixed models. Wiley-

Interscience, New York, USA, 325 p.

Morales, M. 2015. Sciplot Package: Scientific Graphing Functions for Factorial Designs. R

package version 1-1.0. 6p. Available from (https://cran.r-

project.org/web/packages/sciplot/sciplot.pdf) (Consulted: 20/10/2015).

Miina, J., Hotanen, J.-P., and Salo, K., 2009. Modelling the abundance and temporal variation

in the production of bilberry (Vaccinium myrtillus L.) in Finnish mineral soil forest. Silva

Fennica 43 (4): 577–593.

Moran, E. V., Bewick, S., and Cobbold, C. A. 2013. Effects of plant genotype and insect

dispersal rate on the population dynamics of a forest pest. Ecology. 94(12): 2792–2802.

Moreau, G., Eveleigh, E. S., Lucarotti, C. J., and Quiring, D. T. 2006. Ecosystem alteration

modifies the relative strengths of bottom-up and top-down forces in an herbivore population.

Journal of Animal Ecology. 75: 853-861.

44

Morimoto, N., & Nakamura, G. 1989. Studies on the Alpine form of the European pine

sawfly, Neodirion sertifer Geoffroy (Hym. Diprionidae) in the central Japanese Alps. Applied

Entomology and Zoology, 24(4):358-371.

Morrison, M. L., Ralph, C. J., Verner, J., and Jehl, J. R. 1990. Avian Foraging: Theory,

Methodology and Applications. Proceedings of an International Symposium of the Cooper

Ornithological Society. California, USA. 515p.

Nageleisen, L.M. & Bouget, C., coord., 2009. « Forest insect studies: methods and

techniques. Key considerations for standardisation. An overview of the reflections of the «

Entomological Forest Inventories » working group (Inv.Ent.For) ». Les Dossiers Forestiers

n°19, Office National des Forêts, 144 p.

Neuvonen, S., & Niemalä, P. 1991. Do differences in inducible resistance explain the

population dynamics of birch and pine defoliators? In: Baranchiko, Y. N., Mateon, W. J.,

Hain, F. P., and Payne, T. L. (eds.). Forest Insect Guilds: Patterns of Interaction with Host

Trees.. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. NE-153. 103-112.

Nevalainen, S., Sirkiä, S., Peltoniemi, S. and Neuvonen, S. 2015. Vulnerability to pine sawfly

damage decreases with site fertility but the opposite is true with Scleroderris canker damage;

results from Finnish ICP Forests and NFI data. Annals of Forest Science. 42(7):909-917

Niemelä, P., Tuomi, J., and Lojander, T. 1991. Defoliation of the Scots Pine and Performance

of Diprionid Sawflies. Journal of Animal Ecology, 60 (2): 683-692.

Niemelä, P., Chapin S. F., Danell, K., and Bryant J. P. 2001. Herbivory-mediated responses of

selected Boreal forests to climate change. Climatic Change 48: 427–440.

Oksanen, L., Fretwell, S. D., Arruda, J., and Niemelä, P. 1981. Exploitation Ecosystems in

Gradients of Primary Productivity. The American Naturalist. 118(2):240-261.

Oksanen, L., and Oksanen, T. 2000. The Logic and Realism of the Hypothesis of Exploitation

Ecosystems. The American Naturalist. 155(6): 703-723.

Olofsson, E. 1987. Mortality factors in a population of Neodiprion sertifer (Hymenoptera:

Diprionidae). Oikos, 48: 297–303.

Olofsson, E. 1989. Transmission of the nuclear polyhedrosis virus of the European pine

sawfly from adult to offspring. Journal of Invertebrate Pathology. 54:322-330.

Ovaskainen, O., & Hansky, I. 2002. Transient Dynamics in Metapopulation Response to

Perturbation. Theoretical Population Biology 61: 285–295.

Pasquier-Barre, F., Gerí, C., Goussard, F., Auger-Rozenberg, M. A., and Grenier, S. 2000.

Oviposition preference and larval survival of Diprion pini on Scots pine clones in relation to

foliage characteristics. Agricultural and Forest Entomology. 2: 185-192.

Power, M. E. 1992. Top-down and Bottom-up forces in food webs: Do plants have primacy?

Ecology. 73:733-746.

Price, P. W., Roininen, H., and Ohgushi, T. 2005. Adaptive Radiation into Ecological Niches

with Eruptive Dynamics: A Comparison of Tenthredinid and Diprionid Sawflies. Journal of

Animal Ecology.74 (3):397-408.

Pschorn-Walcher, H. 1987. Interspecific competition between the principal larval parasitoids

of the pine sawfly, Neodiprion sertifer (Geoff.) (Hym: Diprionidae). Oecologia. 73: 621-625.

Quinn, G. P. & Keough, M. J. 2002. Experimental Design and Data Analysis for Biologists.

Cambridge University Press The Edinburgh Building, Cambridge, United Kingdom. 557pp

45

Raymond, B., Vanbergen, A., Watt, A., Hartley, S. E., Cory, J. S., and Hails, R. S. 2002.

Escape from pupal predation as a potential cause of outbreaks of the winter moth,

Operophtera brumata. Oikos. 98:219-228.

R Development Core Team, 2013. R: a language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. Available from <http://www.r-

project.org/>.

Rodríguez, A., & Kouki, J. 2015. Emulating natural disturbance in forest management

enhances pollination services for dominant Vaccinium shrubs in boreal pine-dominated

forests. Forest Ecology and Management 350: 1–12.

Rosenheim, J. A. 1998. Higher-Order Predators and the regulation of insect herbivore

populations. Annual Review of Entomology. 43:421-447.

Saikkonen, K., & Neuvonen, S. 1993. European pine sawfly and microbial interactions

mediated by the host plant. In: Wagner, M., & Raffa, K., (eds). Sawfly life history

Adaptations to woody plants. Academic Press, Inc. UK. 431-445.

Sharov, A. A. 1993. Biology and population dynamics of the common pine sawfly Diprion

pini L., in Russia. In: Teoksessa & Haack (eds.). Sawfly life History Adaptations to Woody

Plants. Academic Press, Inc. 409-429.

Stenberg, J. A. 2015. Outbreaking herbivore escapes parasitoid by attaining only a small body

size. Ecosphere. 6(2):21. http://dx.doi.org/10.1890/ES14-00378.1

Sullivan, T. P., Sullivan, D. S., Reid, D. G., and Leung, M. C. 2004. Weasels, Voles, and

Trees: Influence of Mustelid Semiochemicals on Vole Populations and Feeding Damage.

Ecological Applications. 14(4):999-1015.

Tanhuanpää, M., Ruohomäki, K., Kaitaniemi, P., and Klemola, T. 1999. Different impact of

pupal predation on populations of Epirrita autumnata (lepidopter; Geometridae) within and

outside the outbreak range. Journal of Animal Ecology. 68:562-570.

Terborgh, J., Lopez, L., Nuñez, P. V., Rao, M., Shahabuddin, G., Orihuela, G., Riveros, M.,

Ascanio, R., Adler, G. H., Lambert, T. D., and Balbas, L. 2001. Ecological Meltdown in

Predator-Free Forest Fragments. Science. 294:1923-1926.

Tonteri, T., Hotanen, J-P., and Kuusipalo, J. 1990a. The Finnish forest site type approach:

ordination and classification studies of mesic forest sites in southern Finland. Vegetatio 87:

85-98.

Tonteri, T., Kari, M., and Tapani, L. 1990b. Compositional gradients in the forest vegetation

of Finland. Journal of Vegetation Science. 1: 691-698.

Tonteri, T. 1994. Species richness of boreal understory forest vegetation in relation to site

type and successional factors. Annals of Zoologica Fennici. 31: 53-60.

Tscharntke, T. Bommarco, R. Clough, Y. Crist, T. O. Kleijn, D Rand, T. A. Tylianakis, J. M.

van Nouhuys, S. & Vidal, S. 2007. Conservation biological control and enemy diversity on a

landscape scale, Biological Control. 43(3): 294–309.

Turkington, R. 2009. Top-down and bottom-up forces in mammalian herbivore-vegetation

systems: an essay review. Botany. 87:723-739.

Venables, W. N., & Ripley, B. D. 2002. Modern Applied Statistic with S. Fourth Edition

Springer. 504p

46

Verdolin, J. L. 2006. Meta-analysis of foraging and predation risk tradeoff in terrestrial

systems. Behavioral Ecology and Sociobiology. 60(4):457-464.

Virtanen, T., Neuvonen, S., Nikula, A., Varama, M., and Niemelä, P. 1996. Climate change

and the risks of Neodiprion sertifer outbreaks on Scots pine. Silva Fennica. 30(2-3): 169-177.

Volney, W. J. A., & Flemming, R. A. 2000. Climate change and impacts of boreal forest

insects. Agriculture, Ecosystems and Environment. 82:283–294.

Wallner, W. E. 1987. Factors affecting insect population dynamics: Differences between

Outbreak and Non-Outbreak Species. Annual Review of Entomology. 32:317-40.

White, T. C.R. 2006. The world is green. Journal of Vegetation Science 17: 539-540.

Yakamura, K., Yokozawa, M., Nishimori, M., Ueda, Y., and Yokosuka, T. 2006 How to

analyze long-term insect population dynamics under climate change: 50-year data of three

insect pests in paddy fields. Population Ecology. 48, 31–48.

47

7. APPENDIX.

Appendix 1. Table with all the data used in the analysis coded by; Obs (observations) Forest

(forest vegetation type M=Mesic, Sx=Sub-xeric and X=xeric), Area (J=Jaamankangas and

K=Kruununkangas), Site (J1, J2, J3 and K1, K2, K3), Season (summer and autumn), Total

predation, Vertebrate predation and Invertebrate predation.

Obs Forest Area Site Season Total

predation

Vertebrate

predation

Invertebrate

predation

1 M J J3 autumn 9 5 4

2 M J J3 autumn 7 4 3

3 M J J3 autumn 6 3 3

4 M J J3 autumn 15 4 11

5 M K K3 autumn 4 1 3

6 M K K3 autumn 14 7 7

7 M K K3 autumn 13 8 5

8 M K K3 autumn 20 10 10

9 M J J3 summer 22 22 0

10 M J J3 summer 10 4 6

11 M J J3 summer 19 18 1

12 M J J3 summer 9 0 9

13 M K K3 summer 9 3 6

14 M K K3 summer 16 15 1

15 M K K3 summer 11 8 3

16 M K K3 summer 7 2 5

17 Sx J J2 autumn 10 3 7

18 Sx J J2 autumn 5 4 1

19 Sx J J2 autumn 7 2 5

20 Sx J J2 autumn 10 3 7

21 Sx K K2 autumn 4 3 1

22 Sx K K2 autumn 14 14 0

23 Sx K K2 autumn 9 1 8

24 Sx K K2 autumn 4 1 3

25 Sx J J2 summer 6 6 0

26 Sx J J2 summer 6 6 0

27 Sx J J2 summer 3 2 1

28 Sx J J2 summer 4 2 2

29 Sx K K2 summer 0 0 0

48

30 Sx K K2 summer 5 4 1

31 Sx K K2 summer 10 9 1

32 Sx K K2 summer 1 0 1

33 X J J1 autumn 5 1 4

34 X J J1 autumn 4 0 4

35 X J J1 autumn 4 0 4

36 X J J1 autumn 3 2 1

37 X K K1 autumn 5 2 3

38 X K K1 autumn 6 1 5

39 X K K1 autumn 6 3 3

40 X K K1 autumn 5 0 5

41 X J J1 summer 2 2 0

42 X J J1 summer 9 3 6

43 X J J1 summer 0 0 0

44 X J J1 summer 0 0 0

45 X K K1 summer 0 0 0

46 X K K1 summer 1 0 1

47 X K K1 summer 6 6 0

48 X K K1 summer 1 1 0

49

Appendix 2. More detail description of the three types of forest vegetation used; Xeric (Fig

1.), Sub-xeric (Fig 2.) and Mesic (Fig 3.)

Xeric or Calluna type:

This forest type corresponds to dry soils and

canopy layer mainly composed by Pinus

sylvestris (Scots pine forests), rarely mixed with

other tree species. Considered by Cajander as dry

and dryish land forest class. The ground

vegetation is dominated by Calluna vulgaris.

Lichen vegetation is nearly always present

forming a continuous cover depending on the

driest heath. Mosses are almost absent, while

herbs and grasses are scarce. Dwarf shrubs are

abundant in general and xerophilous species.

Figure 1. Xeric forest vegetation type (Calluna type)

(Source: Mar Ramos Sanz.)

Sub-xeric or Vaccinum type:

This type is special because the understorey

vegetation of these forests has mixed attributes

from two forest vegetation classes. Thus,

although in general the understorey vegetation of

these forests is mesophilous it has some species

that corresponds to dry and dryish forest class.

These forests hold a mixture of lichens and

mosses. Herbs and grasses can be present at some

degree. Dwarf shrub vegetation mainly consists

of Vaccinum vitis-idaea. Despite this class can

hold other types of tree species the most

abundant specie was P. sylvestris

Figure 2. Sub-xeric forest vegetation type (Vaccinum

type) (Source: Mar Ramos Sanz.)

Mesic or Myrtillus type:

In general the understorey vegetation of these

forests is mesophilous. These forests hold an

abundant cover of mosses (Hylocomium,

Dicranum etc). Lichens are scarce with the

exception of epiphytes. Herbs and grasses can be

present at some degree. In this case there were

moderate to abundant. Dwarf shrub vegetation

mainly consisted of Vaccinum myrtillus. The

humus layer is well developed consisted in dry

peat. In this study, these forests were dominated

by Scots pine mixed with different broadleaved

trees species with almost no presence of Spruce.

Figure 3. Mesic forest vegetation type (Oxalys-

Myrtillus type) (Source: Mar Ramos Sanz.)

50

Appendix 3. Graphical representation of the weather variables (data available at Mekrijärvi

research station http://mekri.uef.fi/saa/2014/Tiedot.htm) and the two rain gauges.

Figure 1. Graphical representation of the rain (mm) for the previous years (2012, 2013) and

the year of the experiment (2014) including the two rain gadgets data (for Kulho and

Jaamankangas) (Above) and mean temperature (ºC) for these five months for 2012, 2013 and

2014 (bellow).

0

20

40

60

80

100

120

140

160

180

May June July August September

Rain (mm)

2012 2013 2014 Kulho 2014 Jaam 2014

0,00

5,00

10,00

15,00

20,00

25,00

May June July August September

Average temperature (ºC)

2012 2013 2014

51

Appendix 4. Graphical representation of the experimental design set in the field for both

areas Kruununkangas (62° 36' 55"N, 29° 55'E) and Jaamankangas (62° 42´N, 29° 43'E).

52

Appendix 5. Graphical representation of the histograms of the response variables.

Figure 2. Histograms for the response variables: Total predation (above left), vertebrate

predation (above right) and invertebrate predation (bellow left).

Histogram of total

total

Fre

qu

en

cy

0 5 10 15 20 25

05

10

15

20

Histogram of vertebrate

vertebrateF

req

ue

ncy

0 5 10 15 20 25

05

10

15

20

25

30

35

Histogram of invertebrate

invertebrate

Fre

qu

en

cy

0 2 4 6 8 10 12

05

10

15

20