Journal of Invertebrate Pathology 99 (2008) 82–86

Contents lists available at ScienceDirect

Journal of Invertebrate Pathology

journal homepage: www.elsevier .com/ locate /y j ipa

The distribution of Paenibacillus larvae spores in adult bees and honey andlarval mortality, following the addition of American foulbrood diseased broodor spore-contaminated honey in honey bee (Apis mellifera) colonies

Anders Lindström a,*, Seppo Korpela b, Ingemar Fries a

a Swedish University of Agricultural Sciences, Department of Ecology, Box 7044, SE-750 07 Uppsala, Swedenb MTT Agrifood Research Finland, FI-31600 Jokioinen, Finland

a r t i c l e i n f o a b s t r a c t

Article history:Received 16 January 2008Accepted 19 June 2008Available online 28 June 2008

Keywords:ApisPaenibacillus larvaeAmerican foulbroodSpore transmissionContaminated honey

0022-2011/$ - see front matter � 2008 Elsevier Inc. Adoi:10.1016/j.jip.2008.06.010

* Corresponding author. Fax: +46 18 672890.E-mail address: anders.lindstrom@ekol.slu.se (A. L

Within colony transmission of Paenibacillus larvae spores was studied by giving spore-contaminatedhoney comb or comb containing 100 larvae killed by American foulbrood to five experimental coloniesrespectively. We registered the impact of the two treatments on P. larvae spore loads in adult bees andhoney and on larval mortality by culturing for spores in samples of adult bees and honey, respectively,and by measuring larval survival. The results demonstrate a direct effect of treatment on spore levelsin adult bees and honey as well as on larval mortality. Colonies treated with dead larvae showed imme-diate high spore levels in adult bee samples, while the colonies treated with contaminated honey showeda comparable spore load but the effect was delayed until the bees started to utilize the honey at the endof the flight season. During the winter there was a build up of spores in the adult bees, which mayincrease the risk for infection in spring. The results confirm that contaminated honey can act as anenvironmental reservoir of P. larvae spores and suggest that less spores may be needed in honey, com-pared to in diseased brood, to produce clinically diseased colonies. The spore load in adult bee sampleswas significantly related to larval mortality but the spore load of honey samples was not.

� 2008 Elsevier Inc. All rights reserved.

1. Introduction

American foulbrood is a common and widespread bacterial(Paenibacillus larvae) disease of honeybee (Apis mellifera) larvae (El-lis and Munn, 2005). It is one of the most severe diseases of honeybees in apiculture and often destroys honey bee colonies if left un-treated once clinical symptoms have emerged (Hansen andBrødsgaard, 1999). The bacterium produces extremely resilientand long-lived spores (Shimanuki, 1997). Individual larvae becomeinfected before they are 24-h old (Hansen and Brødsgaard, 1999),and once infected they will not recuperate. The larvae often diein the capped cell stage approximately 11 days after egg laying(Shimanuki, 1997) but a large proportion of infected larvae mayalso die before cell capping (Genersch et al., 2005). The bacteriawill sporulate in the decaying larva and the remains of the larvawill become characteristically brown and mucilaginous (Shima-nuki, 1997).

It is known that large numbers of spores are required to estab-lish infections in honey bee colonies (Sturtevant, 1932; Park et al.,1939; Thompson and Rothenbuhler, 1957; Hansen and Brødsgaard,1995). It has also been shown that colonies can maintain relatively

ll rights reserved.

indström).

large numbers of spores over several seasons without any clinicalsymptoms of AFB being manifested (Hansen and Rasmussen,1986; Hansen et al., 1988a,b; Hornitzky and Clark, 1991; Steinkr-aus and Morse, 1992; Fries et al., 2006).

Larval remains are exclusively transmitted between colonies bybeekeepers through inadvertent shifting of brood combs contain-ing remains of larvae that have succumbed to AFB. Contaminatedhoney can be transmitted either by beekeepers shifting extractedor unextracted honey combs between colonies or by the beesthemselves through robbing of colonies weakened by AFB or otherreasons. Robbing is considered to be of major importance for thetransmission of AFB (Hansen et al., 1988a,b; Ratnieks, 1992;Goodwin et al., 1993; Shimanuki, 1997; Hansen and Brødsgaard,1999; Fries and Camazine, 2001; Graaf et al., 2001).

Although it is widely recognised that honey will retain spores ofP. larvae (Sturtevant, 1936; Hansen and Rasmussen, 1986), and thatinter-colony transmission of spores in contaminated honey do oc-cur (Hansen et al., 1988a,b; Ratnieks, 1992; Goodwin et al., 1993;Shimanuki, 1997; Hansen and Brødsgaard, 1999; Fries and Cam-azine, 2001; Graaf et al., 2001), the relative importance of contam-inated honey for intra-colony transmission of P. larvae remainsunknown.

Previously, the significance of inter-colony transmission routesof pathogens for the evolution of virulence in honey bee pathogens

A. Lindström et al. / Journal of Invertebrate Pathology 99 (2008) 82–86 83

have been emphasised by Fries and Camazine (2001). However, thesignificance for disease development of intra-colony transmissionroutes has not been investigated. Different main routes of withincolony transmission may influence disease development in the col-ony differently and, thus, affect between colony transmissionopportunities.

In this study, the significance of the within colony source of P.larvae spores for larval mortality and distribution of spores onadult bees and in honey in honey bee colonies is investigated.

2. Materials and methods

The experiment was set up in Southern Finland(N60�43.0150E23�51.0780) during the summer 2006. A total of 13colonies were used, with three colonies left untreated and usedas controls. The colonies were distributed in groups of 1 (one site)to 3 (one site) colonies/site. The distance between sites was at least500 m. The control colonies were all in one site and in the othersites, where more than one colony was situated, colony distanceswas between 2 and 4 m to minimize drifting.

Five colonies were each given a honey comb from a heavily AFBinfected colony (>500 clinically diseased larvae). The combs con-tained approximately 1.9 � 107, 3.2 � 107, 4.5 � 107, 5.6 � 107

and 8.3 � 107 colony forming units (cfu), respectively, based onmicrobial growth (method, see below) and weight of the combs.The honey combs were uncapped and inserted in the super ofthe experimental colonies, adjacent to the 1st brood frame, outsideof it. Another five colonies were given a piece of brood comb con-taining the remains of approximately 100 larvae killed by AFB.

According to literature, one dead larva contains approximately2.5 � 109 spores (Sturtevant, 1932). To test this, we counted thespores from 15 dead larvae taken from the colony from whichthe dead brood was taken. The counting was done in a Helber bac-teria counting chamber. The results showed that the larvae con-tained a mean of 2.8 � 109 (SD 6.5 � 108) spores. However, thenumber of cfu based on microbial growth (method, see below)was 4.6 � 106, only 0.2% of the total number of spores. The growthrate of spores taken from bacterial colonies grown on MYPGP agarplates is only between 1% and 10% (Forsgren, Stevanovic et al.,2007). Thus, these colonies received approximately 4.6 � 108 cfuper colony, roughly 10 times more than the colonies given honeycomb. The comb pieces containing the dead larvae were fitted toa brood comb from the centre of the brood room of the experimen-tal colonies. To fit the comb piece in the brood comb, a piece ofequal size was cut away and the new piece fitted into the hole.The colonies were placed in a forest environment where foragingplants are not abundant. However, the colonies maintained a broodnest with 7–10 Langstroth frames throughout the season.

The rationale behind the different number of spores given to theexperimental groups is that the cfu levels recorded in the honeycombs have previously been shown to cause clinical AFB symp-toms in honey bee colonies fed such honey (Sturtevant, 1932; Parket al., 1939; Thompson and Rothenbuhler, 1957; Hansen andBrødsgaard, 1995). Similarly, when 50–100 clinically diseased lar-vae are visible in honey bee colonies, the infection is likely to pro-duce further clinical disease symptoms (Hansen and Brødsgaard,1999). Thus, the experiment is not set up to compare infectiondevelopment in colonies given the same number of P. larvae sporesin comb material or in honey, but rather to compare an infectionpressure from two different sources, with potential of producingclinical disease in treated colonies.

With weekly intervals for 5 weeks post-inoculation of spores,500 individual brood cells with eggs were marked and followeduntil cell capping in all 13 colonies. Two days prior to cell marking,the queens were placed in an excluder cage with an empty comb

suitable for egg laying to provide eggs of same age. The combwas placed in the centre of the brood nest. After marking the combwas placed in a box with brood combs separated by a queen exclu-der from the rest of the brood area, so that the queen could not laymore eggs into the empty cells, If a cell had contained an egg, butwas empty upon subsequent inspection it was regarded as an act ofhygienic behaviour from the adult honey bees. Sheets of transpar-ent plastic film were used to mark and keep track of the eggs. Thefilm was secured over the comb and five areas containing 100 cellswere outlined on the film. Every empty cell and cells containingeggs were indicated on the film, as were capped cells or cells con-taining older larvae, honey or pollen. When the comb was re-in-spected the film was secured in same way and the combcompared with the image on the plastic film. When brood becamescarce in the end of August, registrations of brood mortality wereterminated.

Adult bees were sampled once a week for the first 4 weeks andthen once a month until May the following year. The bees wereshaken from combs in the brood nest. Concurrently with samplinga visual inspection of the brood frames was performed to find anyclinical symptoms. If the weather was too cold to open the hive inthe winter, dead bees from the bottom board were sampled. Adultbee samples consisted of >100 adult bees from each colony thatwere shaken or brushed into a plastic bag. The bags were sealedand samples put in a deep freezer and stored until cultivation.Samples of adult honey bees were cultured following the protocolby Lindström and Fries (2005).

Honey samples were taken from honey combs in the supersonce a week for 4 weeks post-infection and then once a monthuntil September, when sampling was terminated. To minimizethe effect of eventual uneven spore distribution in the honeycombs, samples were taken from the comb by scooping withthe edge of a sample jar along the surface of the comb, from dif-ferent parts of the comb, sampling approximately 100 g honeyinto the jar. In the colonies where contaminated honey combshad been inserted, the samples were taken from other honeycombs. The jars were heat treated in 40 �C to float, and separate,the wax from the honey. The honey was then mixed with a glassrod and a 5 g subsample was taken. To cultivate samples, the 5 gsubsample from the collected honey was mixed with 3.5 ml ofwater and heat treated in 88 �C for 15 min to reduce contamina-tion. The suspension was plated onto MYPGP agar with 3 lgnalidixic acid/ml using a 10 ll inoculation loop. Plates wereincubated for 7 days in 36 �C with 5% CO2. The bacterial colonieswere counted manually.

A genetic analysis of AFB spores from the experimental honey-combs and from the diseased brood, used to infect colonies con-firmed that the isolates were both of the ERIC I type (Generschet al., 2005). Both honey combs and the infected brood used inthe experiment came from the same AFB diseased colony.

For the statistical analysis of spore loads in adult bee samplesand honey samples we used a generalized mixed model (the GLIM-MIX procedure) in the SAS Institute statistical software (version9.1.3, 2007). As class variable treatment and time was used andthe cfu was used as response variable. Colony was used as a ran-dom term. Post hoc analysis of significant effects was calculatedwithin the procedure using least squares means test. Analysis ofmortality data was conducted in the same way.

3. Results

Clinical symptoms were not recorded in any colony the firstseason of the experiment, as observed in the weekly inspections.The following spring, two colonies that were given dead brooddeveloped clinical disease symptoms and were shaken onto

84 A. Lindström et al. / Journal of Invertebrate Pathology 99 (2008) 82–86

foundation. Also one colony that was given contaminated combdeveloped weaker symptoms, constantly less than 10 visible dis-eased cells during the second season of the experiment. One colonythat was given a contaminated honey frame died in the spring andwas still symptomless for foulbrood. This colony had shown thehighest mean cfu value both in honey and in the adult bee samplesof all the colonies in the experiment in the first season. No honey orbee samples from the control colonies were positive for P. larvae.

Spore loads of P. larvae in adult bee samples differed signifi-cantly between the two treatments over the whole period(P < 0.00001) but this difference is mainly caused by the large ini-tial difference recorded (Fig. 1). In Fig. 1, in which the P. larvaespore loads of the adult bee samples from the experimental colo-nies have been plotted against time, adult bees from colonies trea-ted with contaminated honey combs have a low P. larvae sporeload for four consecutive weeks when suddenly the spore loadequals the spore load from the colonies treated with deceased lar-vae in brood comb. In Fig. 1, the season can be divided into twoparts—the flight season (July, August, and September) and the win-ter season (October–May). In the flight season there is a significanteffect of treatment (P = 0.0003) but no effect from date(P = 0.7070). From October and onwards, there are no differencesbetween treatments (P = 0.3261) but there is a significant increasein P. larvae spore loads in the adult bee samples (P = 0.0280).

Spore loads of P. larvae in honey samples differed significantlybetween the two treatments (P = 0.0193) considering the wholeexperimental period, although the difference is significant for onesampling occasion only (7/8) when each sampling occasion is ana-lyzed separately. In Fig. 2, the P. larvae spore loads in honey sam-ples from the experimental colonies have been plotted againsttime. Colonies treated with P. larvae contaminated honeycombshave a higher P. larvae spore load than colonies treated with de-ceased brood. The results show a significant effect of treatmenton P. larvae spore loads in honey samples (P = 0.0193) as well asa significant effect of date (P = 0.0489). Since honey was sampledfrom July until September, samples cannot be divided into differentseasons.

The larval mortality for the two treatments and the control isplotted against time in Fig. 3. The generalized mixed model results

24/7 7/

821

/8 4/9

3/10

8/11

8/12

0

1

2

3

4

Log1

0 (c

fu)

D

Fig. 1. Paenibacillus larvae spore loads in adult bee samples from colonies treated with cocolonies are not shown in the figure since all samples were negative. Vertical bars repre

are shown in Table 1. Analysis of data from adult honey bee sam-ples revealed significant effects on larval mortality of treatment(P = 0.0422) and the sample cfu (P = 0.0053). Honey sample analy-sis showed a significant effect from treatment on larval mortality(P = 0.0017), but not from sample cfu.

4. Discussion

Our results strongly suggest that the colony level distribution ofP. larvae spores differ significantly depending on intra-colony ori-gin of spores. At first it appears as if the massive infection pressurefrom 100 AFB scales produces a higher risk for developing clinicaldisease, compared to use of the spore-contaminated honey, basedon spore loads of adult bees (Fig. 1). However, the large differencerecorded between treatments in adult bee spore numbers, dimin-ishes over time (Fig. 1). This result suggests that not only sporedose is important for disease development (Lindström, 2007), butthe source of spores may be equally important in determining col-ony fate. From the presented data it appears that spores in honey,over time will produce similar spore loads on adult bees, as muchhigher spore doses presented in infected brood (Fig. 1). In spite ofthe 10-fold difference in spore dose between treatments, the num-ber of colonies that developed clinical disease did not differ be-tween treatments. Two colonies treated with dead brooddeveloped clinical AFB, whereas one colony given spore-contami-nated honey developed clinical AFB symptoms. This furtheremphasize that spore source is of fundamental importance fordetermining the risk of disease development.

In both treatments, there were increases in spore loads of adultbees during the winter. It has been previously established thatthere is a clear positive correlation between the spore loads ofadult bees and the occurrence of clinical symptoms of AFB(Lindström and Fries, 2005). Thus, the results suggest that theremight be seasonality in disease outbreaks, although this is previ-ously reported not to be the case (Bailey and Ball, 1991). Our datafurther imply that the effect of contaminated honey on adult beespore loads is similar to the effect of AFB-killed brood althoughthe effect is delayed until the flowering season is over or a spellof bad weather force the bees to utilize the honey stores. This re-

15/1

15/2

15/3

15/4

ate

ntaminated honey (dotted line) or deceased larvae in brood comb (full-line). Controlsent the standard deviation.

24/7 31/7 7/8 14/8 21/8 28/8

Date

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Log1

0 (c

fu)

Fig. 2. Paenibacillus larvae spore loads in honey samples from colonies treated with contaminated honey (dotted line) or deceased larvae in brood comb (full-line). Controlcolonies are not shown in the figure since all samples were negative. Vertical bars represent the standard deviation.

0.0

5.0

10.0

15.0

20.0

25.0

30.0

24.7. 31.7. 7.8. 14.8. 21.8.

Date

% m

orta

lity

control

feeding

comb

Fig. 3. The larval mortalities in percent for the two treatments and the control. Vertical bars represent the standard deviation.

Table 1Summary of generalized linear mixed model results for honey bee (Apis mellifera)larval mortality in colonies treated with P. larvae contaminated honey, deceasedbrood or untreated control colonies

Effect Bee samples Honey samples

NumDF

DenDF

F P NumDF

DenDF

F P

Treatment 2 53 3.36 0.0422 2 53 7.23 0.0017Sample

Cfu1 53 8.47 0.0053 1 53 1.32 0.2561

Results for adult bee samples and honey samples are displayed separately since theeffect varies.

A. Lindström et al. / Journal of Invertebrate Pathology 99 (2008) 82–86 85

sult is contradictory to a similar study where 77% of colonies givenP75 scales of American foulbrood and 12% of colonies fed syrupcontaining 5 � 108 spores developed clinical disease (Park et al.,1939). The different outcome in the present experiment may re-flect that this experiment collected samples and made observa-tions over a longer time period and possibly, spore-contaminatedhoney take longer to manifest detectable infection compared todiseased larvae. Differences in spore doses between the experi-ments, in physiological traits (Rothenbuhler and Thompson,

1956), hygienic behaviour (Spivak and Gilliam, 1998) and/or differ-ences in virulence between bacterial strains (Genersch et al., 2005)may also partly explain differences in experimental outcomes.

The results suggest that, at least under Nordic conditions, con-taminated honey may act as an environmental reservoir of P. larvaespores, with spores being disseminated when the bees utilize theirhoney stores. If bees can not defecate outside the hive and rely so-lely on the honey stores for food, there is a build up of fecal mattermixed with P. larvae spores in the rectum of adult bees and suchspores remain viable (Bitner et al., 1972). If bees defecate insidethe hive, spores are released into the hive environment. This prob-ably increases the risk of larval infection due to a more spore-con-taminated environment. However, only three colonies out of theten experimental colonies developed clinical symptoms, suggest-ing that the process leading to clinical disease is more complicatedthan P. larvae spore accumulation in adult bees and possible trans-mission through defecation. The results are congruent with previ-ous studies (Hansen and Rasmussen, 1986; Hansen andBrødsgaard, 1999) where colonies are reported to have spores inhoney for several years without developing clinical disease.

Analysis of the mortality data shows that the treatment effectwas significant for both sample types (adult bees or honey). How-ever, only the spore load in the adult bee samples had a significant

86 A. Lindström et al. / Journal of Invertebrate Pathology 99 (2008) 82–86

relationship with the larval mortality, a relationship that was lack-ing in the honey samples (Table 1). Allover, the mortality level var-ied between 5% and 15% in the treated colonies compared tobetween 2.5% and 4.5% in the non-treated colonies (Fig. 3). Similarmortality levels have been reported previously when feeding colo-nies massive amounts of spores (Thompson and Rothenbuhler,1957). It has previously been demonstrated that analysis of adulthoney bee samples are more sensitive than honey samples whenmonitoring colonies for the presence of P. larvae (Nordströmet al., 2002; Lindström and Fries, 2005). The present study impliesthat the spore load of adult honey bees are more closely linked tothe mortality of the honey bee larvae than the spore load of honeygiving further emphasis to the argument of using sample of adultbees rather than samples of honey, when screening colonies forP. larvae. Although there is an effect on larval mortality from treat-ment compared to untreated controls (Table 1), there is no signif-icant difference between the two treatments, spores in honey ordiseased brood. Thus, infecting colonies through feeding contami-nated honey may require less infectious material, compared toinfecting colonies through inserting infected brood.

None of the experimental colonies showed any clinical symp-toms during the first season of the experiment, although the sporedose varied from approximately 107 in the colonies given contam-inated honey to 1011 in the colonies treated with infected broodcomb. The following spring, three colonies became clinically dis-eased, two of which had been given diseased brood and one thathad received contaminated comb. The minimum number of sporesin honey reported to provoke clinical symptoms varies from5 � 106 (Pankiw and Corner, 1966) to 2 � 109 (Hansen and Brødsg-aard, 1995). Thompson and Rothenbuhler (1957) fed colonies5 � 1010 spores daily for 20 days and managed to induce a few clin-ically diseased cells. In 23% of 200 colonies given ‘‘not less than 75scales of American foulbrood” no clinical symptoms developed(Park et al., 1939). It has been reported that colonies showing100 clinically diseased cells or more are not likely to recuperatebut will eventually succumb to the disease (Woodrow, 1943).There are several possible reasons for this huge variation in sporedose needed to manifest clinical symptoms, which makes the out-come of a specific spore dose hard to predict. Individual larvaeshow variation in physiological traits that results in different sus-ceptibility (Rothenbuhler and Thompson, 1956). There is also var-iation in the virulence between strains of the bacteria (Generschet al., 2005). At the colony level, there is variation in the hygienicbehaviour of the adult honey bees (Spivak and Gilliam, 1998). Fur-thermore, ambient factors like nectar availability and nectar flowintensity play a role (Momot and Rothenbuhler, 1971).

Although the data from this study initially can be explained asspore dose dependent (Fig. 1), since the colonies treated with deadlarvae received a higher spore dose, this difference is eliminated bythe end of August, when spore levels on adult bees become similarin both treatments. Therefore it seems unlikely that the resultsfrom this study should depend only on the initial spore dose.

In conclusion, the data obtained in this study strongly suggestthat the within colony source of P. larvae spores is of importancefor the development of spore loads on adult bees and in honeyand, thus, have an impact on the disease development within col-ony. As a consequence, the within colony spore source is alsoimportant for the transmission of spores between colonies. Fur-thermore, our data imply that the spore load of adult bees is closelylinked to the mortality of honey bee larvae, while there is no signif-icant relationship between the spore loads of honey and larvalmortality. This strongly suggests that samples of adult bees have

a higher predictive value tracing risks for development of clinicaldisease symptoms, compared to samples of honey.

References

Bailey, L., Ball, B.V., 1991. Honey Bee Pathology. Academic press, London.Bitner, A.R., Wilson, W.T., Hitchcock, J.D., 1972. Passage of Bacillus larvae spores

from adult queen bees to attendant workers (Apis mellifera). Ann. Entomol. Soc.Am. 65, 899–901.

Ellis, J.D., Munn, P.A., 2005. The worldwide health status of honey bees. Bee World86, 88–101.

Forsgren, E., Stevanovic, J. Fries, I. 2007. Variability in germination and intemperature and storage resistance among Paenibacillus larvae genotypes.Vet. Microb. Published online doi: 10.1016/j.vetmic.2007.12.001.

Fries, I., Camazine, S., 2001. Implications of horizontal and vertical pathogentransmission for honey bee epidemiology. Apidologie 32, 199–214.

Fries, I., Lindström, A., Korpela, S., 2006. Vertical transmission of Americanfoulbrood (Paenibacillus larvae) in honey bees (Apis mellifera). Vet. Microbiol.114, 269–274.

Genersch, E., Ashiralieva, A., Fries, I., 2005. Strain- and genotype-specific differencesin virulence of Paenibacillus larvae subsp. Larvae, a bacterial pathogen causingAmerican foulbrood disease in honey bees. App. Environ. Microbiol. 71, 7551–7555.

Goodwin, R.M., Perry, J.H., Brown, P., 1993. American foulbrood disease part III:spread. The New Zealand Beekeeper 219, 7–10.

Graaf, D.C.d., Vandekerchove, D., Dobbelaere, W., Peeters, J.E., Jacobs, F.J., 2001.Influence of the proximity of American foulbrood cases and apiculturalmanagement on the prevalence of Paenibacillus larvae spores in Belgianhoney. Apidologie 32, 587–599.

Hansen, H., Brødsgaard, C.J., 1995. Field trials with induced infection of Bacilluslarvae. In: XXXIV International Apicultural Congress of Apimondia Lausanne,Apimondia Publishing House; Bucharest, Romania.

Hansen, H., Brødsgaard, C.J., 1999. American foulbrood: a review of its biology,diagnosis and control. Bee World 80, 5–23.

Hansen, H., Rasmussen, B., 1986. The investigation of honey from bee colonies forBacillus larvae. Dan. J. Plant Soil Sci. 90, 81–86.

Hansen, H., Rasmussen, B., Christensen, F., 1988a. A preliminary experimentinvolving induced infection from Bacillus larvae. Dan. J. Plant Soil Sci. 92, 11–15.

Hansen, H., Rasmussen, B., Christensen, F., 1988b. Infection experiments withBacillus larvae. In: XXXII International Apicultural Congress of Apimondia, Riode Janeiro, Apimondia Publishing House.

Hornitzky, M.A.Z., Clark, S., 1991. Culture of Bacillus larvae from bulk honey samplesfor the detection of American foulbrood. J. Apicult. Res. 30, 13–16.

Lindström, A. 2007. Distribution of Paenibacillus larvae spores among adult honeybees (Apis mellifera) and the relationship with clinical symptoms of Americanfoulbrood. Microb. Ecol. Published online doi: 10.1007/s00248-007-9342-y.

Lindström, A., Fries, I., 2005. Sampling of adult bees for detection of Americanfoulbrood (Paenibacillus larvae subsp larvae) spores in honey bee (Apis mellifera)colonies. J. Apicult. Res. 44, 82–86.

Momot, J.P., Rothenbuhler, W.C., 1971. Behaviour genetics of nest cleaning in honeybees VI. Interactions of age and genotype of bees, and nectar flow. J. Apicult. Res.10, 11–21.

Nordström, S., Forsgren, E., Fries, I., 2002. Comparative diagnosis of Americanfoulbrood using samples of adult honey bees and honey. J. Apicult. Sci. 46, 5–12.

Pankiw, P., Corner, J., 1966. Transmission of American foulbrood by package bees. J.Apicult. Res. 5, 99–101.

Park, O.W., Pellett, F.C., Paddock, F.W., 1939. Results of Iowa’s 1937–1938 honeybeedisease resistance program. Am. Bee J. 79, 577–582.

Ratnieks, F.L.W., 1992. American Foulbrood: the spread and control of an importantdisease of the honey bee. Bee World 73, 177–191.

Rothenbuhler, W.C., Thompson, V.C., 1956. Resistance to American foulbrood inhoney bees. I. Differential survival of larvae of different genetic lines. J. Econ.Entomol. 49, 470–475.

Shimanuki, H., 1997. Bacteria. In: Morse, R.A., Flottum, K. (Eds.), Honey Bee Pests,Predators, and Diseases. A.I. Root Company, Medina, Ohio, USA, pp. 35–54.

Spivak, M., Gilliam, M., 1998. Hygienic behaviour of honey bees and its applicationfor control of brood diseases and varroa. Part I. Hygienic behaviour andresistance to American foulbrood. Bee World 79, 124–134.

Steinkraus, K.H., Morse, R.A., 1992. American foulbrood incidence in some US andCanadian honeys. Apidologie 23, 497–501.

Sturtevant, A.P., 1932. Relation of commercial honey to the spread of Americanfoulbrood. J. Agricult. Res. 45, 257–285.

Sturtevant, A.P., 1936. Quantitative demonstration of the presence of spores ofBacillus larvae in honey contaminated by contact with American foulbrood. J.Agricult. Res. 52, 697–704.

Thompson, V.C., Rothenbuhler, W.C., 1957. Resistance to American foulbrood inhoney bees. II. Differential protection of larvae by adults of different geneticlines. J. Econ. Entom. 50, 731–737.

Woodrow, A.W., States, H.J., 1943. Removal of diseased brood in colonies infectedwith AFB. Am. Bee J. 81, 22–26.