Effects of bud phenology and foliage chemistry of balsam fir and white spruce trees on the efficacy of Bacillus thuringiensis against the spruce budworm, Choristoneura fumiferana

Effects of bud phenology and foliage chemistry of balsam fir andwhite spruce trees on the efficacy of Bacillus thuringiensis againstthe spruce budworm, Choristoneura fumiferana

Nathalie Carisey, Eric Bauce, Alain Dupont* and Sylvain Miron

Faculte de Foresterie et de Geomatique, CRBF, Universite Laval, Sainte-Foy, Quebec, G1K 7P4, Canada and *Societe de Protection des Forets

contre les Insectes et Maladies (SOPFIM), 1780 rue Semple, Quebec, G1N 4B8

Abstract 1 Efficacy of commercial formulations of Bacillus thuringiensis ssp. kurstaki (Btk)against spruce budworm Choristoneura fumiferana was investigated in mixedbalsam fir-white spruce stands. Btk treatments were scheduled to coincide withearly flaring of balsam fir shoots, and later with flaring of white spruce shoots.Btk efficacy on the two host trees was compared and examined according to thefoliar content of nutrients and allelochemicals and the insect developmentalstage at the time of spray.

2 Larvae fed white spruce foliage were less vulnerable to Btk ingestion than larvaefed balsam fir foliage. Higher larval survival on white spruce, observed 10 daysafter spray, was related to higher foliage content in tannins and a lower N/tannins ratio, which might have induced inactivation of Btk toxins.

3 Larval mortality due to Btk did not depend on spruce budworm larval age.4 Foliage protection of both host trees was similar in plots treated with Btk:

larval mortality due to Btk treatment reduced insect grazing pressure on balsamfir trees; meanwhile, suitability of white spruce foliage seemed to decrease veryrapidly, which induced high larval mortality among spruce budworm fed onwhite spruce trees. Nevertheless, following Btk sprays, 50% more foliageremained on white spruce than on balsam fir trees, because of the higherwhite spruce foliage production.

5 Both spray timings achieved similar protection of white spruce trees, but Btktreatments had to be applied as early as possible (i.e. during the flaring ofbalsam fir shoots to optimally protect balsam fir trees in mixed balsam fir-whitespruce stands).

Keywords Bacillus thuringiensis, balsam fir, Btk efficacy, bud phenology,Choristoneura fumiferana, foliage chemistry, foliage production, spray timing,tannins, white spruce.

Introduction

The spruce budworm Choristoneura fumiferana (Clem.) is

an oligophagous tortricid moth, common to the eastern

boreal forests of North America, that feeds primarily on

the current-year foliage of balsam fir, Abies balsamea (L.)

Miller, and white spruce, Picea glauca (Moench) Voss

(Blais, 1983; Sanders, 1991). Boulet (2001) reported that

180millionm3 of timber were lost during the last spruce

budworm outbreak (1967–92) in the Province of Quebec.

Current-year foliage has to be protected against spruce

budworm to prevent extensive defoliation and reduction in

tree volume growth that lead to tree mortality after

4–5 years of severe annual defoliation (Hardy, 1979;

Gagnon & Chabot, 1991).

Forest protection against spruce budworm in the

Province of Quebec (Canada) currently relies on aerial

sprays of the microbial insecticide Bacillus thuringiensis

Berliner ssp. kurstaki (Btk). Btk is a gram-positive

soil bacterium that produces a proteinaceous crystallineCorrespondence: Nathalie Carisey. Tel:þ1 418 656 2131 ext. 13850;

fax:þ1 418 656 3177; e-mail: [email protected]

Agricultural and Forest Entomology (2004) 6, 55–69

# 2004 The Royal Entomological Society

inclusion during sporulation. This inclusion, the d-endotoxin,exhibits highly specific insecticidal activity against Lepi-

doptera (Beegle & Yamamoto, 1992; Dent, 1993). Improve-

ments have been made to Btk formulations, application

technology as well as knowledge of effective droplet size

and lethal dose, but Btk aerial sprays still show variable

efficacy, in part due to several constraints inherent to the

product (van Frankenhuyzen, 1993; van Frankenhuyzen

et al., 1995). First, Btk formulation needs to be ingested by

the larva to be effective. Once activated by the alkaline pH

and proteolytic activity of the insect gut fluids, the toxin

perforates the midgut epithelium, which leads to feeding

cessation and, ultimately, to death through septicemia

(Dent, 1993; van Frankenhuyzen, 1993). Second, foliage on

which the insect feeds needs to be exposed to the Btk aerial

spray, which is why Btk applications do not begin before the

flushing and flaring of new shoots (opened buds with well-

exposed needles). Third, Btk efficacy is potentially affected by

the nutritional and allelochemical composition of food

ingested by the insect. For example, the vulnerability of

gypsy moth (Lymantria dispar L.) and forest tent caterpillar

(Malacosoma distriaHubner) larvae to Btk appear to depend

on host plant species on which the insects grow (Moldenke

et al., 1994; Farrar et al., 1996; Kouassi et al., 2001). Bauce

et al. (2002) reported that spruce budworm larvae that were

fed artificial diet of medium quality were more vulnerable to

Btk than larvae that were fed high quality food. Tannins are

common secondary plant compounds that could reduce the

effectiveness of Btk formulations to L. dispar (Appel &

Schultz, 1994). Moreover, tannins have been shown to

negatively affect the efficacy of Btk d-endotoxins to both

Pieris brassicae L. (Luthy et al., 1985) andHeliothis virescens

Fabricius (Navon et al., 1993). Other chemicals have been

found to enhance Bt sp. effects. Cholorogenic acid and

polyphenol oxidase increase the toxicity ofBtk toHelicoverpa

armigeraHubner (Ludlum et al., 1991). Simple phenols, such

as gallic acid and resorcinol, have also been shown to increase

the activity of Bt ssp. galleriae endotoxin againstH. armigera

(Sivamani et al., 1992).

Most of Canada’s boreal forests are mixed coniferous

stands in which host trees, such as balsam fir and white

spruce, display considerable variation in phenotype, bud

phenology, foliage production and vulnerability to spruce

budworm (Blais, 1976; Hardy, 1979;MacLean&MacKinnon,

1997). Because of the complexity of Btk mode of action, the

challenge of optimally protecting these stands relies on an

understanding of the tri-trophic relationships among Btk,

spruce budworm and host trees. Budbreak in white spruce

occurs 1–4 days later than in balsam fir (Blais, 1957;

Greenbank, 1963). Although balsam fir needles flare quite

rapidly after budbreak, the protective bud scales of white

spruce persist on the tips of the expanding shoots for a long

time following budbreak (Dimond, 1985; Volney & Cerezke,

1992). Thus, the appropriate timing ofBtk applicationsmay be

later for white spruce than for balsam fir (Dimond, 1985). The

first objective of this study was to determine which Btk spray

timing, based on the flaring of balsam fir and white spruce

needles, optimally protects mixed coniferous stands. The

second objective was to compare Btk efficacy between the

two host trees based on foliar nutritional and allelochemical

composition. Lysyk (1989) demonstrated that phenological

development of spruce budworm larvae was different depend-

ing on whether they fed on balsam fir or white spruce trees,

whereas van Frankenhuyzen et al. (1997) showed that lethal

dose requirements per larva increased with larval stage. To

dissociate the influence of foliage quality from that of larval

stage on Btk efficacy, a laboratory experiment using foliage

from the field was conducted. The aims of this study were (i) to

examine Btk efficacy at different larval instars (third, fourth,

fifth and sixth instars) relative to host trees; (ii) to determine

the ingested amounts of Btk-contaminated foliage; and (iii) to

define relationships between Btk mortality rate and amounts

of nutrients (N, P, K, Ca, Mg, total soluble sugars) and

allelochemical compounds (total tannins, total phenolics)

ingested per larva.

Materials and methods

Field experiment

Site description. The experimental site is located in the

Ottawa River Valley of Quebec, Canada (between 45�380-46�010N and 75�330 to 76�330W; mean altitude, 165m asl).

Spruce budworm populations have been at epidemic levels

in the mixed wood stands of this region since 1992. Twenty-

four plots were selected according to the following criteria:

high second-instar spruce budworm populations in autumn;

more than 30% of the basal area in balsam fir and white

spruce; 30–50-year-old spruce-fir stands; moderate previous

defoliation; and easy site access. Plot areas averaged 25 ha.

In each plot, 12 balsam fir and 12 white spruce codominant

trees were randomly selected along transects perpendicular

to the flight lines of the aircraft applying Btk formulations.

Experimental design. Three factors were studied: (i) the

scheduling of the first Btk formulation application that cor-

responded to the moment when buds of balsam fir and white

spruce trees reached the fourth phenological bud stage

(described in the next paragraph); (ii) Btk application rates

that consisted of 0, 1 or 2 applications (with 8–13 days apart)

of 30 billion international units per hectare (BIU/ha); and (iii)

host trees, balsam fir vs. white spruce. Each interaction ofBtk

schedule�Btk application rate, was replicated in four ran-

domly allocated plots. The experimental design is summar-

ized in Table 1.

Bud phenology. The condition of balsam fir buds and

shoots was described according to the scheme of Auger

(Dorais & Kettela, 1982; Juneau, 1989): Stage 1, no appar-

ent bud development (bud in winter condition); Stage 2,

buds swelling and 10–35% of the needles visible; Stage 3,

budbreak, all needles visible but not flaring; Stage 4, needles

flushing and flaring and shoot elongation initiated; Stage 5,

shoots are supple and undergoing elongation.

The phenological stages of white spruce buds and shoots

were adapted from our own observations and from schemes

used by Volney & Cerezke (1992) and Lawrence et al.

(1997): Stage 1, no apparent bud development (bud in

winter condition); Stage 2, buds swelling without separation

of bud scales; Stage 3a, green buds with expanding and

56 N. Carisey et al.

# 2004 The Royal Entomological Society, Agricultural and Forest Entomology, 6, 55–69

thinning scales, but still intact; Stage 3b, all needles are

visible with bud cap separated from the twig base (bud-

break); Stage 4, needles flaring with or without bud cap;

Stage 5, shoot elongation.

Btk formulation applications. Foray 76B, a Btk strain

HD-1 commercial formulation at nominal potency of

20.0BIU/L [Abbott Laboratories (Chicago, IL) on behalf

of Valent Bio-Sciences Corporation (Libertyville, IL)] was

applied to the test plots, described above. Two aircraft were

used, a Piper-Pawnee PA-25 and a Dromader M-18, which

were equipped with six and eight Micronair atomizers

(Micronair Sprayers Ltd, Bromyard, U.K.), respectively.

Micronair atomizers, spinning at 3195 g, were located

within 75% of the total wingspan. The Piper-Pawnee and

Table 1 Summary of the experimental design for the field experiment

Factor 1 Factor 2 Factor 3

Btk schedule No of Btk applications of 30BIU/ha Date of Btk sprays No of plots Host tree

Fourth phenological bud 0 – 1 12 balsam fir

stage of balsam fir tree 12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 28–29 May 2000 1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

2 28–29 May 2000 1 12 balsam fir

7-11 June 2000 12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

Fourth phenological bud

stage of white spruce tree

0 – 1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 3–4 June 2000 1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

2 3–4 June 2000 1 12 balsam fir

11�12 June 2000 12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

1 12 balsam fir

12 white spruce

Btk efficacy against spruce budworm in mixed coniferous stand 57


theDromaderwere flownat 161km/hand193km/h,with30-m

and50-m spraywidths, respectively.Aircraft spraying occurred

early in the morning or at dusk under good weather conditions

(wind speed¼ 6km/h; no rain). The flow rate through the

nozzles was calibrated to deliver 1.5L/ha.

Evaluation of treatment efficacy. A 45-cm branch tip was

collected in the midcrown of the 12 balsam fir and 12 white

spruce trees in each plot on three occasions: 24 h before the

first application (Day 0), 10 days after the first application

(just before the second spray in plots that received two

successive Btk applications; day 10) and at the time when

85% of the larvae had reached the pupal stage (pupal stage

day). Numbers of spruce budworm larvae and their larval

instars (McGugan, 1954) were recorded per branch; an

insect development index (IDI) was calculated according

to the methods of Dorais & Kettela (1982) and Juneau

(1989) where:

IDI ¼ [(no. second-instar larvae � 2) þ (no. third-instar

larvae � 3) þ (no. fourth-instar larvae � 4) þ (no. fifth-

instar larvae � 5) þ (no. sixth-instar larvae � 6) þ (no.

pupae larvae � 7) þ (no. moths � 8)]/(total larvae/branch)

Mortality rates observed on day 10 and pupal stage day

corresponded to larval density reductions between day 0

and day 10 and pupal stage day, respectively.

The phenological bud and shoot stage per branch was

recorded on day 0 and a bud development index (BDI) was

calculated according to the methods of Dorais & Kettela

(1982):

For balsam fir tree, BDI ¼ [(no. first � stage

buds � 1) þ (no. second � stage buds � 2) þ (no. third �stage buds � 3) þ (no. fourth � stage buds � 4) þ (no.

fifth � stage buds � 5)]/(total buds/branch)

For white spruce tree, BDI ¼ [(no. first�stage

buds � 1) þ (no. second�stage buds � 2) þ (no. third (a)

�stage buds � 3) þ (no. third (b)�stage buds � 3.5) þ (no.

fourth�stage buds � 4) þ (no. fifth�stage buds � 5)]/(total

buds/branch)

Bud and shoot defoliation was estimated by the method

of Fettes (1950) (Dorais & Hardy, 1976; Sanders, 1980) on

day 0, day 10 and pupal stage day.

The length and number of new shoots were recorded for

branches collected on pupal stage day. Three shoots per tree

(three trees per host tree and plot) were used to estimate

(i) the number of needles per centimeter of shoot (no.

needles/cm); (ii) the mean dry weight of one needle (mg/

needle); (iii) the dry weight of one shoot axis (mg/shoot

axis); and (iv) the length of one shoot axis (cm/shoot axis).

Foliage was dried in an oven at 70 �C for 3 days. The mean

dry weight of shoot per centimetre (mg/cm) was calculated as:

[(no. needles/cm) � (mg/needle)] þ [(mg/shoot axis)/(cm/

shoot axis)]

Subsequently, to calculate the total dry weight of current-

year foliage produced per branch, the mean dry weight of

shoot per centimetre was multiplied by the mean shoot

length and number of shoots per branch.

Statistical analysis. Bud development index (BDI), insect

development index (IDI) and larval density data on day 0

were subjected to analyses of variance (PROC GLM; SAS

Institute Inc., 1988) in a three-stage crossed nested design.

Four replicate plots were nested within the Btk schedule

(two timings) and Btk application rate (three treatment

levels) factors. The host tree factor was crossed with plots

nested within the Btk schedule and Btk application rate

factors. Larval mortality, defoliation and foliage produc-

tion data were submitted to analyses of covariance with the

initial larval density as a covariate in the previously

described design and are summarized in Table 2. The aver-

age of each parameter, measured on a tree, was firstly

calculated per each host tree and plot. The LSMEANS state-

ment (SAS Institute Inc., 1988), performed for each effect

and interaction, computed least-squares means and multiple

comparisons (least significant difference; LSD), with Bon-

ferroni adjustment for the P-value and confidence limits for

the differences of least squares-means (Bonferroni method

of pairwise comparisons where the adjusted P-value corres-

ponds to the traditional P-value of 0.05 divided by the

number of pairwise comparisons; SAS Institute Inc., 1988).

Foliar bioassays in laboratory

Foliage treatment. Two plots were selected according to

criteria described in the first experiment. The first plot had

Table 2 Sources of variation of the analysis of covariance in the three-stage crossed nested design used for the field experiment

Sources of variation Degrees of freedom

Initial larval density (covariate) 1

Btk schedule (a�1)¼ 1

Btk application rate (b� 1)¼ 2

Btk schedule�Btk application rate (a�1)(b� 1)¼2

Error (a)

Plots� (Btk schedule�Btk application rate) (r�1)ab¼ 3�2� 3¼18

Host tree (c�1)¼ 1

Host tree�Btk schedule (c�1)(a� 1)¼ 1

Host tree�Btk application rate (c�1)(b� 1)¼2

Host tree�Btk schedule�Btk application rate (c�1)(a� 1)(b� 1)¼ 2

Error (b)

Plots�Host tree� (Btk schedule�Btk application rate) (r�1)(c� 1)ab – 1¼ 17

Total rabc – 1¼ 47



no Btk treatment whereas the second was sprayed with

30BIU/ha on 3 June 2000, after white spruce needles had

flushed and flared. The Btk strain HD-1 commercial for-

mulation was Foray 48B (Abbott Laboratories on behalf of

Valent Bio-Sciences Corporation) at nominal potency of

12.7 BIU/L and delivered at a rate of 2.37L/ha. In each

plot, five balsam fir and five white spruce trees were ran-

domly selected. In the treated plot, trees were situated along

lines perpendicular to the flight direction of the aircraft, a

Piper-Pawnee PA-25.

Pre-spray foliage collection and laboratory insect rearing. On

20 May 2000, 25 mid-crown 45-cm branches per host

tree were harvested from different trees in the two selected

plots to sample and rear spruce budworm larvae in the

laboratory; foliage from each tree species was collected

every two days.

A total of 1000 spruce budworm larvae per each host tree

were sampled from branches and reared in 180mL clear

plastic containers (15 larvae per container) filled with either

balsam fir or white spruce buds. A dampened piece of filter

paper was placed on the bottom of the container. Foliage in

each container was replaced every 2 days with fresh material.

For the experiment with postspray foliage (described below),

similar numbers of larvae per each instar (third, fourth, fifth

and sixth) were required at the same moment. Thus, it was

necessary to slow down or increase larval development, and

clear plastic containers filled with larvae were randomly

distributed in four different thermo-electric coolers

(KoolatronsTM, Rochester, U.S.A.) set up at 10, 15, 20 and

25 �C, respectively, for 10–12 days prior to use in the experi-

ment with postspray foliage. These different temperature

treatments (cold or warm conditions) could not have affected

the further results of the experiment, because, first, spruce

budworm larvae in the field are well-adapted to undergo

periods of cold or warm temperatures and, second, larvae

fed on the two host trees were reared in the same thermo-

electric cooler that corresponded to one specific instar.

Postspray foliage collection and laboratory insect rear-

ing. Four hours after the Btk application, five branches

per tree were harvested in the five balsam fir and white

spruce trees previously selected. The current-year shoots

from the five branches per tree were clipped, stored in

plastic boxes and taken to the laboratory where resident

spruce budworm larvae were carefully removed from shoots

to avoid tearing needles and crushing Btk droplets, and

were used as described below.

First, the shoots were used to estimate the foliar water

content. Five shoots per tree were weighed, freeze-dried and

weighed again.

Second, the shoots were used to measure the number of

needles per shoot, the mean dry weight of one needle and

the needle and shoot areas. One hundred needles collected

on five shoots from each tree were freeze-dried and weighed.

Moreover, the projected area of each needle was estimated

using a Digital Image Analysis System (WinFoliaTM,

Regent Instruments Inc., Canada) after placing 50 needles

sampled from each tree on a white sheet. Balsam fir and

white spruce needles have two and four sides, respectively.

Thus, the mean projected area of a needle, for balsam fir

and white spruce was multiplied by 2 and 4, respectively, to

obtain the appropriate needle surface areas. The mean leaf

area of one shoot (mm2) was calculated by multiplying the

needle mean area by the mean number of needles per shoot.

Third, the shoots were used to analyse for N, P, K, Ca,

Mg, total soluble sugar, total tannin and total phenolic

contents, according to methods described by Bauce et al.

(1994) and Bauce (1996). Total tannin concentrations were

quantified by the radial diffusion method (Dement &

Mooney, 1974; Wisdom et al., 1987), which relies on the

formation of binding complexes between tannins and

bovine serum albumin (BSA) protein in agar medium

(Hagerman, 1987). A known quantity of foliage extract is

deposited in a well, cut into agar containing BSA, and the

area of radial diffusion is measured. Thus, total tannin

concentrations are expressed in cm2/mg dry weight foliage.

Fourth, the shoots were used to quantify the Btk toxins

deposited on foliage using the Abbott Deposit Assessment

Method (ADAMKITTM

). This kit, developed on behalf of

Abbott Laboratories by Agdia Inc. (Elkhart, U.S.A.), is

based on the enzyme linked immunosorbent assay (DAS

ELISA) method (van Frankenhuyzen et al., 1998). Five

shoots per tree were placed together in a 50-mL milk sample

bag in which 5mL of solubilization buffer (0.05M trisodium

phosphate, pH12.0, PBS-TA) were added and incubated at

room temperature for 1 h. After agitation, a 1-mL aliquot

was transferred to a 2-mL microcentrifuge tube and an

equal volume of neutralization buffer (0.05M monosodium

phosphate, pH2.2, PBS-TN) was added to obtain a final

pH of 7.5 (van Frankenhuyzen et al., 1998). Because the

foliar samples were too concentrated in Btk toxins to be

directly submitted to the ELISA research kit, they were

diluted with a solution comprising half trisodium phosphate

(pH12.0, PBS-TA) and half monosodium phosphate (pH2.2,

PBS-TN) buffers. For the standard curve, samples of Btk

commercial formulation were collected in the aircraft tank

after spraying, and Btk deposit was expressed in terms of

IU/shoot. Then, the mean quantity of IU deposited on

foliage surface (IU/mm2) per each tree was calculated by

dividing the mean IU/shoot by the mean leaf area of one

shoot (mm2).

Five, the shoots were used to feed the insects and study

the influence of the two host trees on the effect of Btk on

spruce budworm larvae. Shoots collected from each tree

were randomly distributed in four 180mL clear plastic

containers. Fresh shoots in each cup were weighed before

larvae were added. The four containers per host tree

received 15 third-, fourth-, fifth- and sixth-instar larvae,

respectively. The containers were incubated for 5 days at

25 �C, after which the number of dead and surviving larvae

were counted; the remaining foliage was dried (oven at

70 �C for 3 days) and weighed to determine foliage

ingestion. Five days before the Btk treatment, similar

manipulations were carried out with foliage collected in

the control plot. This interval of 5 days between the two

treatments was due to bad weather conditions that delayed

the Btk aerial spray.



Quantity of foliage ingested per larva. Initial fresh foliage

weight, dry weight of the remaining foliage, foliar water

content and mean needle dry weight were used to determine

the number of needles ingested per container (IN), and the

mean number of needles ingested per larva (IN15) when IN

was divided by the 15 spruce budworm larvae. The amounts

of each nutritional (N, P, K, Ca, Mg and total soluble

sugars) and allelochemical (total phenols and tannins) com-

pounds ingested per larva were calculated by multiplying

the nutritional and allelochemical compound concentration

by the amount of ingested foliage dry weight per larva.

Total needles Ingested/container ¼ IN ¼ [([(fresh foliage

weight on day 0a) � (100 � foliar water content)]/100) �remaining foliage dry weight on day 5a)]/mean needle dry

weight

Mean number of needles ingested/larva tested ¼ IN15

¼ IN/15; where 15 ¼ number of larvae per container at the

beginning of the test

Nitrogen consumed (mg)/larva tested ¼ [([(fresh foliage

weight on day 0a) � (100 � foliar water content)]/100 �remaining foliage dry weight on day 5a)/15] � N%; idem

for P, K, Ca, Mg, soluble sugars and phenols consumed

Tannins consumed (cm2)/larva tested ¼ [([(fresh foliage

weight on day 0a) � (100 � foliar water content)]/100 �remaining foliage dry weight on day 5a)/15] � tannins

(cm2/mg)b

[aper container of 15 larvae; bsee earlier for the method to

quantify tannins in foliage]

Statistical analyses. Mean number of needles ingested/

larva tested (IN15) (Btk�treated and control larvae) data

were subjected to an analysis of variance (PROC GLM; SAS

Institute Inc., 1988) in a three-stage crossed nested design

with trees (5 trees) nested within the Btk treatment (control

and Foray 48B) and host tree (two species) factors. The

larval instar factor (four instars) was crossed with trees

nested within the Btk treatment and host tree factors.

IN15 data were ln-transformed to meet assumptions of

normality and homogeneity of variances. The LSMEANS

statement (SAS Institute Inc., 1988), performed for each

effect and interaction, computed least-squares means and

multiple comparisons (Bonferroni method of pairwise com-

parisons).

Mortality rate

No larva died in the control containers during the assay

period and therefore these control treatments were excluded

from the analysis of variance. The mortality data were

subjected to an analysis of variance with a two-stage crossed

nested design with tree samples (five trees) nested within the

host tree factor (two species) (error term for the host tree

factor) and the larval instar factor (four instars) crossed

with trees and nested within the host tree factor (error

term for the larval intar effect and the host tree–larval

intar interaction).

Mean number of needles ingested/larva tested (IN15)

(Btk-treated larvae) data were ln-transformed and subjected

to the previously described analysis of variance with a

two-stage crossed nested design. Transformed data were

normally distributed and homoskedastic.

Mean leaf area of one shoot and IU/mm2 data were

subjected to an analysis of variance in a complete rando-

mized design with host tree as factor and using Btk-treated

trees as experimental units. Data had normal distribution

and homogeneous variances.

Nutritional and allelochemical compound contents were

measured for the two host trees from the control and treated

plots. Due to a 5-day delay between the control and Btk-

treated foliage collection, the analysis of these data indicated

changes in foliar compound contents in time. Thus, data

were subjected to an analysis of variance in a 2� 2 factorial

design with host tree factor (balsam fir and white spruce) and

collection day factor (29 May vs. 3 June) using trees as

experimental units.

Relationships between mortality rate and quantities of

nutritional and allelochemical ingested compounds per

tested larva (independent variables) were analysed by linear

regression using stepwise method at P< 0.1 (PROC STEPWISE;

SAS Institute Inc., 1988). Separate regression analyses were

performed on fifth- and sixth-instar data sets, respectively.

Results

Field experiment

Bud development index (BDI) at Btk application

timing. Six days separated the two Btk timings and within

this period, differences between balsam fir and white spruce

phenological shoot stages were nearly halved. At the balsam

fir Btk spray timing, balsam fir shoots were 14% more

developed than white spruce shoots. At the white spruce

Btk spray timing, they were just 8% more developed than

white spruce shoots (Tables 3 and 4). The first Btk applica-

tion was delayed 4–5 days due to bad weather conditions,

and explains why the Btk sprays were carried out when

balsam fir and white spruce shoots had exceeded the stage

4 (Table 4). Nevertheless, this delay could not significantly

affect conclusion on Btk timing: by the time, most balsam

fir shoots were at stage 4, the white spruce bud development

index was equal to 3.68 (data not shown). This index was

very close to that of white spruce (3.98) when Btk sprays

timed on balsam fir schedule were conducted (Table 4). This

small difference in shoot development is unlikely to have

affected the number of Btk droplets per area unit.

Insect development index (IDI) at Btk application

timing. At the time of balsam fir Btk sprays, spruce bud-

worm larvae fed on balsam fir or white spruce trees showed

similar larval development. Six days later, at the time of the

white spruce Btk sprays, larvae fed on white spruce trees

appeared to have grown more rapidly than on balsam fir

trees (Tables 3 and 4).

Initial larval density. Larval densities (larvae per 45-cm

branch tip) evaluated on balsam fir trees were similar at the

time of balsam fir and white spruce Btk sprays, but larval

density on white spruce increased from the balsam fir Btk

spray timing to the white spruce Btk spray timing (Tables 3

and 4). During their larval development, spruce budworm



larvae tended to move vertically as well as horizontally

within trees, as reported previously by Regniere et al.

(1989). Because the initial larval density (i.e. density one

day before the first Btk sprays) depended on host tree and

Btk schedule, which could affect larval mortality and defo-

liation, initial larval density was used as a covariate for the

analyses of covariances of the different parameters mea-

sured.

Larval mortality. Larval mortality (%) observed 10 days

after the first Btk application was significantly influenced by

the Btk application rate, the effect of which depended on the

Btk schedule (Table 5). In the controls, larval mortality rate

observed on day 10 was higher at the white spruce Btk spray

timing than at the balsam fir Btk spray timing (Table 5;

Fig. 1), which is consistent with literature that reported

increase in spruce budworm natural mortality during the

season (Royama, 1984; Regniere & You, 1991). Btk sprays,

carried out at the balsam fir timing, induced higher larval

mortality than those carried out at the white spruce timing

(Fig. 1). Larval mortality, observed on day 10, depended on

both host tree and Btk application rate factors (Table 5). On

this day, differences in larval mortality between balsam fir

and white spruce trees were 42% and 64%, for the control

and Btk-treated plots (at this time, all treated plots had

received only one Btk application), respectively (Fig. 2a).

On average, larval mortality in Btk-treated plots was

much higher on balsam fir trees (approximately 66%) than

on white spruce trees (around 24%) (Fig. 2a). In fact, Btk

application rate did not influence significantly larval mor-

tality from day 0 to day 10 on white spruce trees (Fig. 2a).

Btk spray timing did not influence larval mortality from

day 0 to pupal stage day (Table 5), but mortality during this

period was higher on white spruce control trees than on

balsam fir control trees (Table 5; Fig. 2b). By contrast, one

or two Btk applications induced higher larval mortality on

balsam fir trees (90%) than on white spruce trees (84%)

(Fig. 2b).

Defoliation. Analysis of covariance indicated that initial

defoliation (day 0) was higher in control plots than in Btk-

treated plots (Tables 6 and 7). The choice of plots in the

Ottawa River Valley was subjected to logistic constraints;

some owners did not allow SOPFIM to conduct Btk sprays

on their forests, which then were used as control plots.

Although the result was not statistically significant, these

control plots had 62% higher larval population densities

than plots dedicated to the Btk treatment (Table 3), and

this might have influenced the results of the initial defolia-

tion. Pre-spray defoliation was nearly two-fold higher at

white spruce Btk spray timing than at balsam fir Btk spray

timing (Tables 6 and 7).

Btk schedule and Btk application rate factors signifi-

cantly influenced defoliation observed on day 10 (Table 6).

Table 3 Summaries of the analyses of variance, in a three-stage crossed nested design, on bud development index (BDI), insect development

index (IDI) and larval density measured at the beginning of the field experiment (day 0)

BDI (Day 0) IDI (day 0)

Larval density

(larvae/45-cm branch tip) (day 0)

Sources of variation d.f. P MSE P MSE P MSE

Btk schedule 1 0.0005 0.0001 0.64

Btk application rate 2 0.74 0.13 0.07

Btk schedule�Btk application rate 2 0.95 0.99 0.66

Mean square error (a) 18 0.0593 0.0335 801.76

Host tree 1 0.0001 0.0001 0.0001

Host tree�Btk schedule 1 0.008 0.008 0.005

Host tree�Btk application rate 2 0.53 0.89 0.22

Host tree�Btk schedule�Btk application rate 2 0.37 0.4 0.44

Mean square error (b) 18 0.0141 0.0056 68.37

d.f., P-values for F-tests and the mean square errors (MSE) used for the F-tests (in italic). Numbers in bold indicate statistically significant effects

at P< 0.05. When interactions are significant, the corresponding simple effects are not in bold type.

Table 4 Statistically significant interaction (P< 0.05) between Btk schedule (Balsam fir, BF) and white spruce (WS) timing, and host tree on

foliage and larval population characteristics measured on day 0 in the field experiment.

Btk schedule� host tree

Day 0 BF timing�BF BF timing�WS WS timing�BF WS timing�WS

Bud development index 4.55�0.07b 3.98� 0.07d 4.75� 0.07a 4.38�0.07c

Insect development index 4.21�0.04c 4.27� 0.04c 4.89� 0.04b 5.07�0.04a

Larval density 35�5b,c 42� 5b 32� 5c 54�5a

(larvae/45-cm branch tip)

Mean�2SE followed by the same letter do not differ significantly according to the Bonferroni pairwise comparison method.



Defoliation on day 10 was 247% higher in the control plots

than that in the Btk-treated plots (Table 7), and this defoli-

ation was nearly two-fold higher at white spruce timing

than at balsam fir timing (Table 6). Defoliation from day 0

to day 10 increased only in control plots.

At the end of larval development, Btk application rate

resulted in different levels of foliage protection against

spruce budworm, depending on the host trees (Table 6).

Although defoliation of balsam fir and white spruce trees

was similar in plots sprayed with one or two Btk applica-

tions, defoliation in control plots was higher on balsam fir

trees than on white spruce trees (Fig. 3a). Defoliation

observed on pupal stage day was 133% higher in the control

plots than that in plots treated with one or two Btk sprays

(Fig. 3a). Moreover, at the pupal stage day, defoliation on

balsam fir trees was higher when Btk sprays were carried

out at white spruce timing instead of balsam fir timing.

Defoliation on white spruce trees was not influenced by

the Btk schedule (Table 6; Fig. 4).

Table 5 Summaries of the analyses of covariance, in a three-stage crossed nested design, on larval mortality measured on day 10 and pupal

stage day

Mortality (%) 10days

after the first Btk spray

Mortality (%) from the first Btk

spray (day 0) to pupal stage day

Sources of variation d.f. P MSE P MSE

Initial larval density (covariate) 1 0.0019 0.0044

Btk schedule 1 0.66 0.64

Btk application rate 2 0.016 0.08

Btk schedule�Btk application rate 2 0.011 0.32

Mean square error (a) 18 192.88 53.92

Host tree 1 0.0002 0.47

Host tree�Btk schedule 1 0.95 0.08

Host tree�Btk application rate 2 .0016 0.0001

Host tree�Btk schedule�Btk application rate 2 0.53 0.69

Mean square error (b) 17 125.35 14.01

d.f., P-values for F-tests and the mean square errors (MSE) used for the F-tests (in italic). Numbers in bold indicate statistically significant effects

at P< 0.05. When interactions are significant, the corresponding simple effects are not in bold type.

Figure 1 Larval mortality (%) on both host trees 10days after the first

Btk spray, according to the three Btk application rates (control,

1�30BIU, and 2�30BIU, respectively, 0, 1, and 2 applications of

30BIU/ha) and the Btk schedule based on bud flaring stage of either

balsam fir (BF) or white spruce (WS).

100

90

80

70

60

50

40

30

20

10

0Control 1×30BIU 2×30BIU

±2SE

Balsam firWhite spruce

(a)

Larv

al m

orta

lity

(%)

from

Day

0 to

Day

10

100

90

80

70

60

50

40

30

20

10


±2SE

(b)

Larv

al m

orta

lity

(%)

from

Day

0 to

Pup

al S

atge

Day

Figure 2 (a) Larval mortality (%) 10 days after the first Btk spray and

(b) larval mortality (%) at pupal stage day depending on three Btk

application rates and the host trees.



Table

6Summariesoftheanalysesofcovariance,in

athree-stagecrossednesteddesign,ondefoliation(%

)measuredonday0,day10andpupalstagedayandonremainingfoliage(g/45-cm

branchtip)

Defoliation(%

)onday0

Defoliation(%

)onday10

Defoliation(%

)atpupalstageday

New

foliageproduction

(g/45-cm

branchtip)

Remainingnew

foliage

(g/45-cm

branchtip)

Sourcesofvariation

d.f.

PMSE

PMSE

PMSE

PMSE

PMSE

Initiallarvaldensity(covariate)

1<0.0001

<0.0001

<0.0001

–<0.0001

Btk

schedule

10.0018

<0.0001

0.4

0.13

0.11

Btk

applicationrate

20.03

0.0002

0.001

0.006

0.03

Btk

schedule

�Btk

applicationrate

20.39

0.08

0.37

0.5

0.38

Meansquare

error(a)

86

18

0.0057

0.0138

224.6

213.86

130.56

Hosttree

10.64

0.3

0.68

<0.0001

0.02

Hosttree�Btk

schedule

10.8

0.09

0.049

0.21

0.52

Hosttree�Btk

applicationrate

20.23

0.11

0.0003

0.041

0.49

Hosttree�Btk

schedule�

Btk

applicationrate

20.23

0.09

0.96

0.5

0.62

Meansquare

error(b)

17

0.0024

0.0034

49.2

63.32

46.45

d.f.,P-valuesforF-testsandthemeansquare

errors

(MSE)use

dfortheF-tests(inita

lic).Ananalysisofvariancewasperform

edonnewfoliageproductiondata

because

theinitiallarvald

ensity

usedas

covariate

wasnotstatistically

significant.Numbers

inbold

indicate

statistic

ally

significanteffects

atP<0.05.Wheninteractio

nsare

significant,thecorresp

ondingsim

ple

effects

are

notin

bold

type.

Table

7Statistically

significanteffects

(atP<0.05)ofBtk

schedule,Btk

applicationrate

orhosttreeondefoliation(%

)measuredonday0andday10andonamounts

ofremainingfoliage

(g/45-cm

branchtip)atpupalstageday

Btk

schedule

Btk

applicationrate

Hosttree

BFtiming

WStiming

Control

1�30BIU

2�30BIU

Balsam

fir

Whitespruce

Defoliationonday0

12�4b

21�4a

26�6a

13�5b

12�5b

––

Defoliationonday10

22�5b

40�5a

59�11a

17�8b

17�7b

––

Remainingfoliageonpupalstageday

––

6.57�11.08b

24.35�7.76a,b

29.27�7.48a

15.41�4.09b

24.71�3.99a

Mean�2SEfollowedbythesameletterwithin

thesamerowperfactordonotdiffersignificantlyaccordingto

theBonferronip

airwisecomparisonmethod.Horizontalb

ars

indicate

thattheeffectof

this

factorwasnotstatistically

significant.BF,Balsam

fir;WS,whitespruce.



Foliage production (g) and remaining foliage (g) on a 45-

cm branch tip (in dry weight). Btk treatment of one or two

applications resulted in 161% higher white spruce foliage

production relative to the controls. Balsam fir foliage pro-

duction was not significantly affected by Btk application

rate and was similar to white spruce foliage production in

control plots (Table 6; Fig. 3b). Overall, foliage production

was 38% higher on white spruce than on balsam fir.

Foliage remaining on 45-cm branch tips was significantly

affected by Btk application rate (Table 6). Some 270% and

345% more foliage remained in plots treated with one and

two Btk applications, respectively, relative to the control

plots (Table 7). There was significantly more foliage (60%)

left on white spruce trees than on balsam fir trees (Tables 6

and 7).

Foliar bioassays in laboratory

The mean leaf area of a white spruce shoot (2860mm2) was

twice as large as the mean leaf area of a balsam fir shoot

(1421mm2) (F1,9¼ 19.86; P¼ 0.0021). However, Btk

deposition on balsam fir foliage (0.42� 0.32 IU/mm2;

mean� 2 SE) was similar to that on white spruce foliage

(0.56� 0.32 IU/mm2; mean� 2 SE) (F1,9¼ 0.03; P¼ 0.87).

Number of needles ingested per larva tested (IN15) (Btk-

treated and control larvae.) The number of needles ingested

per larva tested (IN15) was significantly influenced by Btk

treatment (control vs. Foray 48B), the effect of which

depended on host tree ingested by the larva (F1,16¼ 19.17;

P¼ 0.0005). Larvae reared on foliage from the control plot

consumed the same amount of needles regardless of host

trees (approximately 84 needles). However, larvae reared on

Btk-treated balsam fir ingested seven-fold fewer needles

(five needles) than those reared on Btk-treated white spruce

(35 needles). A three-way interaction between Btk treat-

ment, host tree and instar was statistically significant

(F3,48¼ 3.57; P¼ 0.0207). The amounts of needles ingested

per larva increased with larval instar in the control (balsam

fir and white spruce foliage), and in the Btk treatment with

white spruce foliage. However, larval instar did not influ-

ence the ingestion of larvae reared on Btk-treated balsam fir

foliage.

Mortality rate. The effect of Btk on mortality depended

on the type of host trees (F1,8¼ 16.25; P¼ 0.0038). Mortal-

ity rates among larvae reared on balsam fir or white spruce

foliage were 85 and 43% (� 15%, �2 SE), respectively.

Spruce budworm larval instar did not influence the effect

of Btk (F3,24¼ 1.87; P¼ 0.16).

Number of ingested needles per tested larva (IN15) (Btk-

treated larvae). The effect of host trees on the number of

needles ingested per larva tested (IN15) was influenced by

larval instar (F3,24¼ 6.77; P¼ 0.0018). Third-, fourth-, fifth-

and sixth-instar larvae reared on balsam fir ingested similar

numbers of needles as did third- and fourth-instar larvae fed

on white spruce foliage. However, fifth- and sixth-instar

larvae fed on white spruce foliage consumed significantly

greater number of needles (59% more) than younger larvae

fed white spruce foliage and larvae of all intars fed balsam

fir foliage (Fig. 5).

Nutritional and allelochemical compound contents. Host

tree and collection day did not influence nitrogen, phos-

phorous and magnesium contents of the foliage collected

(Table 8). Balsam fir foliage contained 40% more calcium

100

90

80

70

60

50

40

30

20

10


±2SE


(a)

Def

olia

tion

(%)

on P

upal

Sta

ge D

ay

55

50

45

40

35

30

25

20

15

10

0

5

Control 1×30BIU 2×30BIU

±2SE

(b)

Fol

iage

pro

duct

ion

(g/4

5-cm

bra

nch

tip)

Figure 3 (a) Defoliation (%) and (b) new foliage production (g/45-cm

branch tip) observed on pupal stage day depending on three Btk

application rates and the host trees.

100

90

80

70

60

50

40

30

20

10

0BF timing WS timing


±2SE

Def

olia

tion

(%)

on P

upal

Sta

ge D

ay

Figure 4 Defoliation (%) observed on pupal stage day according to

the Btk schedule based on bud flaring stage of either balsam fir

or white spruce and the host trees of either balsam fir or white

spruce.



than white spruce foliage, but calcium content in foliage did

not change over 5 days in either species (Table 8). Total

soluble sugars in either host trees were significantly lower

on 29 May than on 3 June, and there was a host tree effect;

white spruce foliage contained 29% more sugars than bal-

sam fir foliage (Table 8). Potassium levels increased 12%with time from 29May to 3 June (Table 8). There were more

phenols in white spruce foliage on 29 May than 3 June, but

the opposite was true for balsam fir foliage (Table 8). White

spruce foliage contained 19% more total tannins than

balsam fir foliage, and tannin contents were higher on 29

May than on 3 June in both host trees (Table 8). The ratio

of N to tannins, which is an index of the amounts of nitro-

gen not bound to tannins in foliage (Feeny, 1969), was 33%higher in balsam fir foliage than in white spruce foliage

(Table 8), and was higher on 29 May than on 3 June in

both host trees.

Relationships between mortality rate and ingested amounts

of nutrient and allelochemical compounds. Stepwise regres-

sions on mortality rate vs. ingested quantities of nutritional

and allelochemical compounds per larva tested produced

two significant equations for the fifth and sixth instars,

respectively:

Fifth instar: mortality rate (%) ¼ � 3.14 � (tannins

ingested/larva tested) þ 91.17; (r2 ¼ 70%); the standard

errors of the slope and intercept estimates were 9.6 and

0.72, respectively.

Sixth instar: mortality rate (%) ¼ � 11.39 � (tannins

ingested/larva tested) þ 550.18 � (K tested/ingested larva) þ85.47; (r2 ¼ 87%); tannins ingested/larva tested: partial

r2 ¼ 69%; K ingested/larva tested: partial r2 ¼ 18%9 the stand-ard errors of the slope (tannins ingested and K ingested/tested

larva) and intercept estimates were 3.3, 182 and 6.7, respectively.

These results indicated that mortality among fifth- and

sixth-instar larvae that had ingested Btk, decreased with the

amount of ingested tannins. In sixth-intar larvae, the

amounts of ingested K had positive effects on Btk mortality

rate.

Discussion

Experiments in the field and laboratory clearly showed that

Btk caused lower mortality when spruce budworms fed on

white spruce foliage compared to balsam fir foliage. Ten

days after field application of Foray 76B, larval mortality

was 2.8-fold higher on fir than on spruce trees. Moreover,

Btk applications on white spruce trees did not significantly

increase mortality among budworm population beyond

untreated controls. Under warm and constant temperature

(25 �C), larval mortality due to Foray 48B ingestion was

higher in the laboratory than in the field, but mortality

among larvae fed balsam fir foliage was also two-fold

higher than that among larvae fed white spruce foliage.

This response could not be linked to differences in the

quantity of Btk toxins deposited on foliage. When Btk

formulation was applied at the flaring of balsam fir buds,

white spruce buds were still closed, and Btk droplets prob-

ably did not reach the insect target. However, at the flaring

of white spruce, young shoots of both host trees were fully

opened to receive a uniform mist of Btk spray. Results of

the Btk toxin dosage experiment also indicated that Btk

deposits were similar on balsam fir and white spruce foliage.

Thus, other factors must have influenced the efficacy of Btk

on spruce budworm fed balsam fir compared to white

spruce foliage.

Larvae fed white spruce developed faster than larvae fed

balsam fir and were particularly advanced when Btk was

applied at white spruce Btk timing, which was consistent

with reports in the literature (Blais, 1957; Webb &McLeod,

1957; Lavallee & Hardy, 1988; Lysyk, 1989). However, the

laboratory experiment showed that differences in Btk

efficacy between the two host trees could not be attributed

to differences in larval development. Third, fourth, fifth and

sixth instars fed balsam fir foliage ingested similar amounts

of Btk-contaminated needles and there was no indication

that they differed in mortality (mean¼ 85%). Although

ingestion of Btk-contaminated needles increased with larval

instar (from 12 to 58 needles/larva for third–fourth and

fifth–sixth instars, respectively) on white spruce, larvae fed

white spruce foliage had similar mortality rate (43%)

regardless of larval instar. The increase in food consump-

tion as larvae grew resulted undoubtedly in ingestion of

higher Btk doses, which is consistent with the findings of

van Frankenhuyzen et al. (1997) for spruce budworm larvae

fed artificial diet, that ‘the 3.5-fold increase in lethal dose

requirement per larva from the fourth to the sixth instar is

more than compensated for by the 20-fold increase in

potential feeding rates’. In the present study, larvae fed

white spruce foliage ingested between two- and 12-fold

more Btk-contaminated needles than larvae fed balsam fir

foliage and less than 30% of the larvae died in the field. The

results indicate that larvae fed on white spruce had a high

tolerance to Btk and feeding inhibition may have set in just

before they acquired the lethal dose. By contrast, larvae fed

balsam fir foliage rapidly acquired the lethal dose, although

they ingested just a few needles.

At the time spruce budworm larvae were subjected to

Btk treatment, total tannin contents was higher in white

4.5

4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.03rd Instar 4th Instar 5th Instar 6th Instar


±2SE

Ln (

Num

ber

of n

eedl

es in

gest

ed/la

rva

test

ed)

Figure 5 Effect of the host trees on the natural logarithm of the

number of needles ingested per larva according to the larval

instar.



spruce foliage than in balsam fir foliage. Moreover, mor-

tality due to Btk ingestion was negatively correlated to the

amount of tannins ingested by a larva. This effect may be

related to the protein precipitating ability of tannins

(Zucker, 1983), which may have led to inactivation of the

Btk proteinaceous toxin. Appel & Schultz (1994) showed

that the efficacy of Thuricide, another commercial formu-

lation of Btk, towards gypsy moth also depended on host

trees; red oak foliage was more inhibitory to Btk toxins

than chestnut and quaking aspen, probably because of its

higher concentrations of total phenolics and tannins. Lord

& Undeen (1990) found that tannins released by decaying

vegetation in mosquito breeding sites tended to reduce

Bacillus thuringiensis ssp. israelensis efficacy. Luthy et al.

(1985) demonstrated that Btk protoxin could be inacti-

vated by increasing concentrations of hydrolysable gallo-

tannins and they also showed that Btk d-endotoxin was

strongly inactivated by extracts prepared from spruce nee-

dles. Formation of complexes between tannins and pro-

teins depends on the relative concentrations of tannins and

protein, and ‘a threshold level of tannin is required before

any protein precipitation occurs’ (Zucker, 1983). In the

present study, the ratio of N to tannins from white spruce

foliage was significantly lower than that from balsam fir

foliage, which indicates that more tannins were available

to form complexes with protein and Btk toxin in the mid-

gut of larvae fed white spruce than in that of larvae fed

balsam fir.

The midgut pH of lepidopteran larvae is alkaline and, in

many species, tends to reach pH10 (Berenbaum, 1980;

Gringorten et al., 1992), and this creates a favourable

environment for solubilization and digestion of Btk proto-

xin (Dent, 1993; van Frankenhuyzen, 1993). Larval midgut

pH tended to decrease when larvae fed on acidic food

(foliage or artificial diet) (Keating et al., 1990; Bauce et al.,

2002), and even slight reductions in pH to low alkaline pH

(above 8) could decrease the vulnerability of insects to

nuclear polyhedrosis virus (Stiles & Paschkles, 1980;

Keating et al., 1990) and suppress toxicity of the enzyme-

activated d-endotoxin of Btk (Gringorten et al., 1992).

Further investigations should verify if midgut pH of larvae

fed white spruce is lower than that of larvae fed balsam fir,

and whether this plays a major role in the low efficacy

of Btk formulations on white spruce trees. The protein-

precipitating ability of tannins is also favoured at pH

under 8 (Zucker, 1983) and lower alkaline pH in midgut

of larvae fed white spruce foliage could strengthen the

previous hypothesis of the potential role of white spruce

tannins in the inactivation of Btk toxins.

The positive correlation between the amount of potas-

sium ingested by the larva and mortality due to Btk inges-

tion indicated that potassium availability could enhance Btk

efficacy. Knowles (1994) highlighted the major role of

potassium in the destruction of midgut epithelial cells after

Bt-toxins had induced pores in the epithelium. Morris et al.

(1995) reported that a commercial Btk formulation

supplemented with 0.05% of potassium carbonates improved

Btk efficacy against Mamestra configurata Walker fed on

canola (Brassica napus L.or Brassica rapa L.).Table

8Chemicalc

ontents

(mean�2SE)in

balsam

firandwhitesprucefoliagecollectedon29Mayand3JuneFoliagewasusedto

testBtk

efficacyonsprucebudworm

larvaefedtwodifferent

hosttrees

Hosttree

Collectionday

Balsam

fir

Whitespruce

29May

3June

Hosttree�collectionday

Nutritionalcompounds

N(%

dry

weight)

F1,19¼3.52;p¼0.079

2.38�0.18a

2.15�0.18a

F1,19¼0.82;p¼0.38

2.21�0.18a

2.32�0.18a

F1,19¼0.25;p¼0.62

P(p.p.m

)F1,19¼1.36;P¼0.26

3725�202a

3558�202a

F1,19¼3.69;P¼0.073

3779�202a

3504�202a

F1,19¼2.25;P¼0.15

K(p.p.m

)F1,19¼1.18;P¼0.29

13626�596a

13168�596a

F1,19¼12.34;P¼0.0029

14138�596a

12656�596b

F1,19¼3.43;P¼0.083

Ca(p.p.m

)F1,19¼8.12;P¼0.012

1926�272a

1378�272b

F1,19¼1.20;P¼0.29

1757�272b

1547�272b

F1,19¼0.44;P¼0.52

Mg(p.p.m

)F1,19¼0.14;P¼0.71

1152�66a

1170�66a

F1,19¼2.07;P¼0.17

1194�66a

1127�66a

F1,19¼0.27;P¼0.61

Totalsoluble

sugars

(%dry

weight)

F1,19¼4.87;P¼0.042

5.91�1.1

b7.64�1.1

aF1,19¼5.3;P¼0.035

5.88�1.1

b7.68�1.1

aF1,19¼1.66;P¼0.22

Allelochemicalcompounds

Totalphenols

(%)

F1,19¼0.81;P¼0.38

11.74�1.46a

10.81�1.46a

F1,19¼0.44;P¼0.52

11.62�1.46a

10.94�1.46a

F1,19¼5.62;P¼0.0306

Totaltannins(cm

2/m

gfoliagedry

weight)

F1,19¼14.36;P¼0.0016

0.642�0.046b

0.763�0.046a

F1,19¼19.6;P¼0.0004

0.773�0.046a

0.632�0.046b

F1,19¼1.74;P¼0.21

N/totaltannins

F1,19¼12.21;P¼0.003

3.81�0.38a

2.87�0.38b

F1,19¼9.93;P¼0.0062

2.91�0.38b

3.77�0.38a

F1,19¼0.1;P¼0.75

Means�

2SEfollowedbythesameletterwithin

thesamerow

perfactordonotdiffersignificantlyaccordingto

theBonferronipairwisecomparisonmethod.



In control plots, defoliation was higher on balsam fir

than on white spruce trees at the end of larval development.

Insect grazing pressure was undoubtedly higher on balsam

fir than on white spruce because larval mortality during the

total larval development was lower when larvae fed on

untreated balsam fir foliage than on untreated white spruce

foliage. An increase in raw fibre foliar content (i.e. lignin,

silica, pectin, cellulose and hemicellulose) may be the cause

of this result because it negatively affects larval survival of

spruce budworm (Bauce & Hardy, 1988) as well as larch

budmoth (Zeiraphera diniana (Gn.) (Benz, 1974). The rapid

cessation of white spruce shoot elongation and the start of

shoot lignification have been reported to rapidly make

white spruce foliage less suitable for spruce budworm

nutrition (Thomas, 1987; Lawrence et al., 1997) and these

changes in foliage quality lead to decrease drastically larval

survival. By contrast, balsam fir foliage appears to remain

suitable for a longer period of time than white spruce foliage

because of its slower rate of shoot growth (Greenbank,

1963).

Larvae that ingested sublethal Btk doses exhibit longer

larval development (van Frankenhuyzen et al., 1997; Bauce

et al., 2002). In the Btk-treated plots, the more important

the increase in larval development, the more important are

the risks to encountering unsuitable foliage, especially for

larvae fed on white spruce trees, because of the shoot lig-

nification. Feeding on unsuitable white spruce foliage

tended to decrease larval survival, which led to a reduction

in defoliation of white spruce trees. At the end of the grow-

ing season, defoliation was similar in Btk-treated balsam fir

and white spruce trees, although Btk-related larval mortal-

ity on balsam fir trees was nearly three-fold higher than that

on white spruce trees. White spruce foliage protection was

probably achieved by the combination of a shorter window

of foliage suitability and Btk effects, lethal and sublethal

(feeding inhibition and increase in larval development time).

The present field experiment showed that balsam fir

foliage production was similar in control and Btk-treated

plots. Insect grazing probably compromised shoot elonga-

tion prior to Btk applications, or was able to overtake

balsam fir production (Koller & Leonard, 1981), even

when budworm populations were low and reduced by Btk

treatment (under 16 larvae per 45-cm branch tip). White

spruce shoots can grow longer than balsam fir shoots and

thus, provide more food per unit of larval population

(Greenbank, 1963; Koller & Leonard, 1981; Regniere

et al., 1989). However, this phenomenon was not observed

at high larval densities (42 larvae per 45-cm branch tip);

control balsam fir and white spruce trees showed similar

foliage production, probably because insect grazing by high

larval population densities had outpaced expansion of both

balsam fir and white spruce shoots. On Btk-treated white

spruce, most of the spruce budworm larvae ceased to feed

and defoliation did not increase during the 10 days follow-

ing Btk application (unpublished data). Thus, shoot elonga-

tion was not impeded and foliar production was 60% higher

in white spruce than in balsam fir in Btk-treated plots.

More foliage remained on white spruce than on balsam

fir trees. In control plots, white spruce trees underwent

lower defoliation, because of an increase in budworm mor-

tality at the end of the larval development. In Btk-treated

plots, white spruce and balsam fir trees exhibited similar

levels of defoliation, but 50% more foliage remained on

spruce compared to fir due to the higher foliage production

of the former.

During the last budworm epidemic (1967–92) in the Pro-

vince of Quebec, the main objective of forest protection

against spruce budworm was tree survival. When aerial

spray programmes were planned, one of the decision criteria

was the death risk of host trees. White spruce trees were less

vulnerable to spruce budworm than balsam fir trees, with

spruce dying after 5–7 years of severe successive defoliation

compared to 4–5 years for balsam fir trees (Hardy, 1979;

Gagnon & Chabot, 1991). However, in the present managed

forests, it should be more important to protect stand yield

and more efficient strategies have to be found. The impact

of standard Btk sprays (one or two applications of 30BIU/

ha) on spruce budworm fed on white spruce appeared to

rely more on sublethal effects than on mortality. If the host

tree reduces pest vulnerability to Btk, applications of higher

Btk concentrated formulations than those currently used

could increase the chance that larvae ingest a lethal dose

more rapidly (Appel & Schultz, 1994).

Differences in bud phenology among host trees compli-

cated the scheduling of Btk applications in mixed-coniferous

stands, and the inter-relationship between spruce budworm,

host tree, and Btk at bud flaring needed to be understood.

The present field experiment clearly demonstrated that pro-

tection of mixed balsam fir-white spruce stands, required that

Btk treatment had to be applied as early as possible (i.e. on

the flaring of balsam fir buds). When Btk treatments were

scheduled at the later white spruce flaring, defoliation had

already reached 21% for both tree species, and tree growth

begins to be negatively affected at this level of defoliation

(MacLean, 1985). Furthermore, defoliation of white spruce

at the end of the season was similar at the two Btk timings,

whereas balsam fir defoliation at the end of the season was

higher when Btk was applied at the white spruce flaring

instead of at the balsam fir flaring. The main advantage of

scheduling Btk application at balsam fir bud expansion was

to provide the best protection for this species, and finally the

mixed balsam fir-white spruce stands. These results were in

accordance with those of Cadogan & Scharbach (1993) who

showed that balsam fir defoliation was significantly higher

when stands were sprayed at the flaring of black spruce

(28%) instead of the flaring of balsam fir (18%). They also

found that Btk application timing did not influence the black

spruce final defoliation.

This experiment demonstrates the importance of acquir-

ing a fundamental knowledge of the relationships between

Btk, spruce budworm and host trees (chemical foliage

profile as well as bud phenology) for developing the best

forest protection strategy according to ecological and

economic concerns of the day. Further research of this

kind should be pursued to determine the appropriate and

efficient use of the Btk insecticide in forest protection for this

specific complex (Btk and spruce budworm) in other geograph-

ical locations and for other tree species and defoliators.



Acknowledgements

The authors would like to thank the laboratory and field

team of the forestry department at SOPFIM and at the

Laval University for their helpful technical assistance.

Thanks are addressed to Dan Quiring (Population Ecology

Group, Faculty of Forestry and Environmental Manage-

ment, University of New Brunswick, Fredericton, Canada)

for his critical review of this manuscript. Research funding

was provided by the National Sciences and Engineering

Council of Canada and the Societe de Protection des Forets

contre les Insectes et Maladies.

References

Appel, H.M. & Schultz, J.C. (1994) Oak tannins reduce effective-

ness of Thuricide (Bacillus thuringiensis) in the gypsy moth

(Lepidoptera: Tortricidae). Journal of Economic Entomology, 87,

1736–1742.

Bauce, E. (1996) One and two years impact of commercial thinning

on spruce budworm feeding ecology and host tree foliage

production and chemistry. Forestry Chronicle, 72, 1–6.

Bauce, E. & Hardy, Y. (1988) Effects of drainage and severe

defoliation on the rawfiber content of balsam fir needles and

growth of the spruce budworm (Lepidoptera: Tortricidae).

Environmental Entomology, 17, 671–674.

Bauce, E., Crepin, M. & Carisey, N. (1994) Spruce budworm

growth, development and food utilization on young and old

balsam fir trees. Oecologia, 97, 499–507.

Bauce, E., Bidon, Y. & Berthiaume, R. (2002) Effects of food

nutritive quality and Bacillus thuringiensis on feeding behaviour,

food utilization and larval growth of spruce budworm

Choristoneura fumiferana (Clem.) when exposed as fourth and

sixth-instar larvae. Agricultural and Forest Entomology, 4, 1–14.

Beegle, C.C. & Yamamoto, T. (1992) Invitation paper (C.P.

Alexander Fund): history of Bacillus thuringiensis Berliner

research and development. Canadian Entomologist, 124, 587–616.

Benz, G. (1974) Negative feedback by competition for food and

space and by cyclic induced changes in the nutritional base

as regulatory principles in the population dynamics of the

larch bud moth, Zeiraphera diniana (Gn.) (Lep., Tortricidae).

Zeitschrift fuer Angewandte Entomologie, 76, 196–228.

Berenbaum, M. (1980) Adaptive significance of midgut pH in

larval Lepidoptera. American Naturalist, 115, 138–146.

Blais, J.R. (1957) Some relationships of the spruce budworm,

Choristoneura fumiferana (Clem.) to black spruce, Picea mariana

(Moench) Voss. Forestry Chronicle, 13, 364–372.

Blais, J.R. (1976) Vulnerabilite de l’epinette blanche a la defoliation

par la tordeuse des bourgeons de l’epinette. Revue Bimestrielle de

Recherches, 32, 21.

Blais, J.R. (1983) Trends in the frequency, extent, and severity of

spruce budworm outbreaks in eastern Canada. Canadian Journal

of Forest Research, 13, 539–617.

Boulet, B. (2001) Retrospective-Les enseignements de la derniere

epidemie de tordeuses des bourgeons de l’epinette. Tordeuse des

Bourgeons de l’epinette: l’Apprivoiser dans nos Strategies

d’Amenagement pp. 3–13, Fo18–48/2001F, 27–29 March. Service

Canadien Des Forets, Shawinigan, Quebec, Canada.

Cadogan, B.L. & Scharbach, R.D. (1993) Efficacy of Foray 48B

(Bacillus thuringiensis Berliner) applications against the spruce

budworm, Choristoneura fumiferana (Clemens) (Lepidoptera:

Tortricidae), timed for phenological development of balsam fir

and black spruce. Canadian Entomologist, 125, 479–488.

Dement, W.A. & Mooney, H.A. (1974) Seasonal variation in the

production of tannins and cyanogenic glucosides in the

chaparral shrubs, Heteromeles arbutifolia. Oecologia, 15, 65–76.

Dent, D.R. (1993) The use of Bacillus thuringiensis as an

insecticide. Exploitation of Microorganisms (ed. by D. G. Jones),

pp. 19–43. Chapman & Hall, London.

Dimond, J.B. (1985) Considerations in timing insecticidal sprays

on different host trees of the spruce budworm. Recent advances

in spruce budworms research. CANUSA Spruce Budworms

Research Symposium, pp. 409. Bangor, Maine, 16–20 September

1984. Canadian Forest Service, Ottawa.

Dorais, L.G. & Hardy, Y.J. (1976) Methode d’evaluation de la

protection accordee au sapin baumier par les pulverisations

aeriennes contre la tordeuse des bourgeons de l’epinette.

Canadian Journal of Forest Research, 6, 86–92.

Dorais, L.G. & Kettela, E. (1982) Revue, par region, des techniques

d’inventaire entomologique et d’evaluation des programmes de

pulverisation a grande echelle contre la tordeuse des bourgeons de

l’epinette Choristoneura fumiferana (Clem.). Conseil de l’Est de

la tordeuse des bourgeons de l’epinette. Rapport du comite pour

la standardisation des techniques entomologiques, p. 51.

Ministere de l’Energie et des Resources du Quebec, Quebec,

Canada.

Farrar, R.R.J., Phyllis, A.W.M. & Ridgway, R.L. (1996) Host

plant effects on activity of Bacillus thuringiensis against gypsy

moth (Lepidoptera: Lymantriidae) larvae. Environmental

Entomology, 25, 1215–1223.

Feeny, P.P. (1969) Inhibitory effect of oak tannins on the

hydrolysis of proteins by trypsin. Phytochemistry, 8, 2119–2126.

Fettes, J.J. (1950) Investigations of sampling techniques for

population studies of the spruce budworm on balsam fir in

Ontario. Forest Insect Laboratory, Sault Ste Marie, Annual

Technical Report, 4, 163–401.

Gagnon, R.R. & Chabot, M. (1991) Prevention des pertes de bois

attribuables a la tordeuse des bourgeons de l’epinette, p. 52.

Gouvernement du Quebec, Ministere des forets. Service d’ame-

nagement forestier et service de protection contre les insectes et

les maladies, Quebec.

Greenbank, D.O. (1963) Host trees and the spruce budworm.

Memoirs of the Entomological Society of Canada, 31, 219–223.

Gringorten, J.L., Milne, R.E., Fast, P.G., Sohi, S.S. & Van

Frankenhuyzen, K. (1992) Suppression of Bacillus thuringiensis

d-endotoxin activity by low alkaline pH. Journal of Invertebrate

Pathology, 60, 47–52.

Hagerman, A.E. (1987) Radial diffusion method for determining

tannin in plant extracts. Journal of Chemical Ecology, 13, 437–449.

Hardy, Y.J. (1979) Les recherches sur la tordeuse des bourgeons de

l’epinette. Revue Forestiere Francaise, 31, 217–222.

Juneau, A. (1989) A Review of Aerial Spraying Technology for

Spruce Budworm Control in Private Woodlots in Eastern Quebec,

p. 219. Canadian Forest Service, Ste Foy, Quebec.

Keating, S.T., Schultz, J.C. & Yendol, W.G. (1990) The effect of

diet on gypsy moth (Lymantria dispar) larval midgut pH, and its

relationship with larval susceptibility to a baculovirus. Journal of

Invertebrate Pathology, 56, 317–326.

Knowles, B.H. (1994) Mechanism of action of Bacillus thuringiensis

insecticidal d-endotoxins. Advances in Insect Physiology, 24,

275–308.

Koller, C.N. & Leonard, D.E. (1981) Comparison of energy

budgets for spruce budworm Choristoneura fumiferana

(Clemens) on balsam fir and white spruce. Oecologia, 49, 14–20.

Kouassi, K.C., Lorenzetti, F., Guertin, C., Cabana, J. &Mauffette, Y.

(2001) Variation in the susceptibility of the forest tent caterpillar

(Lepidoptera: Lasiocampidae) to Bacillus thuringiensis variety



kurstaki HD-1: effect of the host plant. Journal of Economic

Entomology, 94, 1135–1141.

Lavallee, R. & Hardy, Y. (1988) Etude en laboratoire du developpe-

ment du Choristoneura fumiferana sur l’Abies balsamea, le Picea

glauca et le Picea rubens. Phytoprotection, 69, 79–86.

Lawrence, R.K., Mattson, W.J. & Haack, R.A. (1997) White

spruce and the spruce budworm: defining the phenological

window of susceptibility. Canadian Entomologist, 129, 291–318.

Lord, J.C. & Undeen, A.H. (1990) Inhibition of the Bacillus

thuringiensis var. israelensis toxin by dissolved tannins. Environ-

mental Entomology, 19, 1547–1551.

Ludlum, C.T., Felton, G.W. & Duffey, S.S. (1991) Plant defenses:

Chlorogenic acid and polyphenol oxidase enhance toxicity of

Bacillus thuringiensis subsp. kurstaki to Heliothis zea. Journal of

Chemical Ecology, 17, 217–237.

Luthy, P., Hofmann, C. & Jaquet, F. (1985) Inactivation of delta-

endotoxin of Bacillus thuringiensis by tannin. FEMS Microbiol-

ogy Letters, 28, 31–33.

Lysyk, T.J. (1989) Stochastic model of eastern spruce budworm

(Lepidoptera: Tortricidae) phenology in white spruce and

balsam fir. Journal of Economic Entomology, 82, 1161–1168.

MacLean, D.A. (1985) Effects of spruce budworm outbreaks on

forest growth and yield. CANUSA Spruce Budworms Research

Symposium, pp. 148–175. Bangor, Maine, 16–20 September

1984. Canadian Forest Service, Ottawa.

MacLean, D.A. & MacKinnon, W.E. (1997) Effects of stand and

site characteristics on susceptibility and vulnerability of balsam

fir and spruce to spruce budworm in New Brunswick. Canadian

Journal of Forest Research, 27, 1859–1871.

McGugan, B.M. (1954) Needle-mining habits and larval instars of

the spruce budworm. Canadian Entomologist, 86, 439–454.

Moldenke, A.F., Berry, R.E., Miller, J.C., Wernz, J.G. & Li, X.H.

(1994) Toxicity of Bacillus thuringiensis subsp. kurstaki to gypsy

moth, Lymantria dispar, fed with alder or douglas fir. Journal of

Invertebrate Pathology, 64, 143–145.

Morris, O.N., Converse, V. & Kanagaratnam, P. (1995) Chemical

additive effects on the efficacy of Bacillus thuringiensis Berliner

subsp. kurstaki against Mamestra configurata (Lepidoptera:

Noctuidae). Journal of Economic Entomology, 88, 815–824.

Navon, A., Hare, J.D. & Federici, B.A. (1993) Interactions among

Heliothis virescens larvae, cotton condensed tannin and the

CryIA (c) d-endotoxin of Bacillus thuringiensis. Journal of

Chemical Ecology, 19, 2485–2499.

Regniere, J. & You, M. (1991) A simulation model of spruce

budworm (Lepidoptera: Tortricidae) feeding on balsam fir and

white spruce. Ecological Modelling, 54, 277–297.

Regniere, J., Lysyk, T.J. & Auger, M. (1989) Population density

estimation of spruce budworm, Choristoneura fumiferana

(Clem.) (Lepidoptera: Tortricidae) on balsam fir and white

spruce from 45-cm mid-crown branch tips. Canadian Entomol-

ogist, 121, 267–281.

Royama, T. (1984) Population dynamics of the spruce budworm

Choristoneura fumiferana. Ecological Monographs, 54, 429–462.

Sanders, C.J. (1980) A Summary of Current Techniques Used for

Sampling Spruce Budworm Populations and Estimating Defoli-

ation in Eastern Canada, p. 33. Environment Canada, Canadian

Forest Service Information Report: O-X-306, Great Lakes

Forest Research Center, Sault Ste Marie, Ontario.

Sanders, C.J. (1991) Biology of North American spruce budworms.

Tortricid Pests, Their Biology, Natural Enemies and Control (ed.

by L. P. S. van der Geest and H. H. Evenhuis), pp. 579–620.

Elsevier Science Publishers, Amsterdam, The Netherlands.

SAS Institute Inc (1988) SAS/STAT User’s Guide, version 6.03.

SAS Institute, Inc., Cary, NC.

Sivamani, E., Rajendran, N., Senrayan, R., Ananthakrishnan, T.N.,

Jayaraman, K. (1992) Influence of some plant phenolics on the

activity of d-endotoxin of Bacillus thuringiensis var. galleriae on

Heliothis armigera. Entomologia Experimentalis et Applicata, 63,

243–248.

Stiles, B. & Paschkles, J.D. (1980) Midgut pH in different instars of

three Aedes mosquito species and the relationship between pH

and susceptibility of larvae to a nuclear polyhedrosis virus.

Journal of Invertebrate Pathology, 35, 58–64.

Thomas, A.W. (1987) The effect of age of current-year shoots of

Picea glauca on survival, development time, and feeding

efficiency of 6th-intar larvae of Choristoneura fumiferana.

Entomologia Experimentalis et Applicata, 43, 251–260.

van Frankenhuyzen, K. (1993) The challenge of Bacillus

thuringiensis. Bacillus Thuringiensis, an Environmental Biopesti-

cide: Theory and Practice (ed. by P. F. Entwistle, J. S. Cory,

M. J. Bailey and S. Higgs), pp. 1–35. John Wiley & Sons, New

York.

van Frankenhuyzen, K., Nystrom, C.W. & Tabashnik, B.E. (1995)

Variation in tolerance to Bacillus thuringiensis among and within

populations of the spruce budworm (Lepidoptera: Tortricidae)

in Ontario. Journal of Economic Entomology, 88, 97–105.

van Frankenhuyzen, K., Gringorten, L., Dedes, J. & Gauthier, D.

(1997) Susceptibility of different instars of the spruce budworm

(Lepidoptera: Tortricidae) to Bacillus thuringiensis var kurstaki

estimated with a droplet-feeding method. Journal of Economic

Entomology, 90, 560–565.

van Frankenhuyzen, K., Nystrom, C. & Dedes, J. (1998) The Use

of the Abbott Deposit Assessment Method and a Foliar Bioassay

to Evaluate Spray Deposition and Treatment Effects Resulting

From Experimental Applications of Foray 48B in the 1997

Gatineau Program. Canadian Forest Service, Great Lakes

Forestry Centre, Sault Ste Marie, Ontario, Canada.

Volney, W.J.A. & Cerezke, H.F. (1992) The phenology of white

spruce and the spruce budworm in northern Alberta. Canadian

Journal of Forest Research, 22, 198–205.

Webb, F.E. & McLeod, J.M. (1957) Studies of Aerial Spraying

Against the Spruce Budworm in New Brunswick IX. Comparative

Studies of the Spruce Budworm on Balsam Fir and Spruce in New

Brunswick, p. 28. Forest Biology Laboratory, Fredericton, New

Brunswick. Department of Agriculture, Science Service, Canada.

Wisdom, C.S., Gonzales-Coloma, A. & Rundel, P.W. (1987)

Ecological tannin assays. Evaluation of proanthocyanidins,

protein binding assays and protein precipitating potential.

Oecologia, 72, 395–401.

Zucker, W.V. (1983) Tannins: Does structure determine function?

An ecological perspective. American Naturalist, 121, 335–365.

Accepted 4 March 2003



Effects of bud phenology and foliage chemistry of balsam fir and white spruce trees on the efficacy of Bacillus thuringiensis against the spruce budworm, Choristoneura fumiferana

Documents