• blackberry rust;
  • host growth;
  • integrated management;
  • invasive plant;
  • leaf spot


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

European blackberry (Rubus fruticosus agg.) is an aggregate of closely related taxa, with at least 15 taxa naturalized in Australia. Biological control of this Weed of National Significance, using the nonindigenous rust fungus Phragmidium violaceum, is effective when the weather is conducive to multiple cycles of infection, but some blackberry taxa escape severe disease. Thirty-one taxa of naturalized R. fruticosus agg. from southeastern Australia were isolated, their DNA phenotype determined and clones of each taxon inoculated with P. violaceum isolate SA1. Disease development was monitored for at least four generations of uredinia on large potted plants under field conditions. Although variation in mean disease severity appeared continuous over the range of Rubus clones tested, counts of uredinia and telia enabled identification of eight resistant taxa. Fine scale variation in susceptibility to rust disease was observed when different clones of R. leucostachys with the same DNA phenotype were found to express either resistance or susceptibility to P. violaceum (SA1). There were significant differences among 23 Rubus taxa rated as susceptible to rust disease in the mean number of leaves emerging per latent period of uredinia (LELPU). Mean LELPU appeared to account for some of the variation in two measures of mean disease severity observed among susceptible Rubus clones, although the correlation was insignificant (0·10 < P > 0·05).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

There are at least 15 taxa of European blackberry (Rubus fruticosus agg.) that have naturalized in Australia following multiple introductions of this woody perennial plant by European settlers in the 19th century (Amor et al., 1998; Evans et al., 2004b). European blackberry, categorized by Thorp & Lynch (2000) as a ‘Weed of National Significance’, infests more than 8 × 106 hectares of pasture, forests and natural ecosystems and is a target for biological control with the nonindigenous rust fungus Phragmidium violaceum, especially in areas that are remote and/or inaccessible to integrated weed management.

Australia has a small subset of the total genetic diversity present in the host in Europe, given that some 300 taxa have been described in central Europe alone (Weber, 1996). The majority of European blackberry taxa are polyploid, facultatively apomictic and pseudogamous (Nybom, 1988). Most seed produced is genetically identical to the mother plant but a proportion of seed may develop following sexual reproduction. Distinguishing species in the genus Rubus is often problematic because morphological diversity is high and interspecific hybridization is common (Alice et al., 2001). The phylogeny of Rubus has been studied by sequencing nuclear ribosomal DNA (Alice & Campbell, 1999), but inadequate taxonomic treatment of the R. fruticosus agg. by morphology has confounded efforts to improve biological control of invasive blackberry genotypes. DNA phenotyping (fingerprinting) has been used to confirm taxonomic determinations in doubtful cases and to support a taxonomic update of nonindigenous Rubus in Australia based on morphology (Evans et al., 2004b). This technique may also be applied to genotype plant material used in pathogenicity studies with P. violaceum, where it can be assumed that if two plants have the same DNA phenotype, they are likely to belong to the same Rubus species, even if they are not necessarily genetically identical.

In their native range (Europe), species of R. fruticosus agg. are commonly infected by P. violaceum. The rust was first reported in Australia in 1984 following an unauthorized introduction (Marks et al., 1984). Evans et al. (2000) reviewed the authorized introduction of P. violaceum isolate F15 to Australia in 1991 for biological control of blackberry. This macrocyclic and autoecious rust fungus (Laundon & Rainbow, 1969) mainly infects leaves, but occasionally also the petioles, green floral organs, unripe fruits and green stems of the plant. All spore states can form under Australian conditions (Washington, 1985).

Disease caused by P. violaceum has reduced the biomass and daughter-plant production of R. polyanthemus in southeastern Victoria (Mahr & Bruzzese, 1998). Rust disease is also having an impact on other susceptible taxa of the R. fruticosus agg. in localities in Victoria where the annual rainfall is greater than 800 mm and the average maximum daily temperature for the month of January is close to 20°C (Pigott et al., 2003). Even when weather conditions were favourable for development of rust disease, one patch of blackberry in Victoria was observed during this study to express severe rust disease, whereas an adjacent patch, presumably a different genotype, had visibly less disease. The existence of both qualitative and quantitative resistance to rust disease might explain the observed difference; for example, ‘gene-for-gene’ interactions are common in both agricultural and natural rust pathosystems (Thompson & Burdon, 1992). Previously, Evans & Bruzzese (2003) characterized leaf age-related disease resistance in R. anglocandicans. In particular, they found that the number of uredinia produced per blackberry shoot was proportional to the number of leaves that were less than 30 days old. They postulated that some differences among blackberry taxa in disease expression might be conferred by variation in rate of leaf emergence and hence the age profile of leaves within the plant canopy at the time of infection by P. violaceum.

The aim of this study was to collect a representative sample of taxa of R. fruticosus agg. in Australia, propagate them clonally and compare rust disease development among clones growing in the same field. Within this overall aim, three hypotheses were tested. Firstly, that a subset of Rubus clones isolated in Australia will be resistant to disease caused by a recent Australian isolate of P. violaceum. Confirmation of the presence of disease-resistant Rubus clones after multiple cycles of uredinial infection will allow validation of methods for controlled-environment bioassay whereby specific host/pathogen interactions are determined after one infection cycle. The second hypothesis proposed that the rate of leaf emergence varies among Rubus genotypes under conditions suitable for both shoot and disease development. While estimates of rates of leaf emergence of various Rubus clones in the absence of disease are of ecological interest, they do not reflect leaf emergence in environments conducive to disease where blackberry patches are often exposed to airborne inoculum of P. violaceum. Disease severity and shoot growth data generated from testing the first two hypotheses were then used in preliminary analyses towards testing a third hypothesis: namely, that the mean rate of leaf emergence for Rubus taxa susceptible to rust disease is correlated with disease severity.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Origin and propagation of plant material

Each Rubus clone originated from a single crown collected from a naturalized infestation of blackberry (Table 1). Techniques for M13/HaeIII DNA phenotyping of Rubus clones are described in Evans et al. (1998, 2000). The majority of Rubus clones were also determined to species level, based on a recent taxonomic update of exotic Rubus in Australia (Evans et al., 2004b). All clones were members of the R. fruticosus aggregate, except clone JH1656, which was R. laudatus, a blackberry of North American origin. Crowns were transported to the Waite Campus of the University of Adelaide and maintained under quarantine conditions as specified by the state of South Australia. Plants were propagated clonally either by taking semihardwood to hardwood cuttings or by daughter-plant production following tip-rooting of vegetative cane apices in potting soil. Cuttings with three nodes were selected from the vegetative shoot. The cutting base was treated with 3 g kg−1 indol-3-butyric acid and planted in 15 cm of sand or a 1:1 sand:peat mixture. After 6–8 weeks, cuttings with roots were planted into a well-drained potting mixture based on pine bark and all plants propagated were maintained in a shade house for approximately 2 years. In the winter prior to bioassay, dormant plants were pruned to ground level.

Table 1.  Origin and DNA phenotype of Rubus clones collected in southeastern Australia between 1996 and 2001
CloneTaxonDNA phenotypeCollection dateCollector(s)OriginaLatitude and longitude (decimal code)Voucher specimen
  • a

    Vic., Victoria; NSW, New South Wales; SA, South Australia; Tas., Tasmania.

  • b

    A blackberry of North American origin.

9607R. anglocandicans32Jul. 1996BruzzeseSomerville, Vic.−38.2385, 145.2355Unavailable
960204R. anglocandicans32Jun. 1996MayoScott Creek, SA.−35.062, 138.6985Unavailable
EB14R. cissburiensis29Dec. 1997Bruzzese, Mahr & EvansGilbert's Road, Strzelecki Ranges, Vic.−38.4748, 146.4207AD99811445
0101R. echinatusNot determinedMar. 2001RudmanNative Plains Road, Mersey Lea, Tas.−41.3333, 146.4667HO505646
EB20R. erythrops25Dec. 1997Bruzzese, Mahr & EvansGellibrand River, Vic.−38.7180, 143.2507AD99811452/3
SR14 (15406)R. laciniatus37Dec. 1996SymonLenswood, SA−34.9167, 138.8333AD99652181/ 99708225
KE1R. laciniatus37Dec. 1997EvansCreswick, Vic.−37.4249, 143.894AD99809203
JH1656R. laudatusbNF3Dec. 1998Hosking & DellowNear Bilpin Primary School, NSW−33.4969, 150.5169AD99851276
JH1566R. leightonii35Feb. 1998HoskingKanagra Boyd National Park, NSW−33.8894, 150.0425AD99810024
JH1660R. leightonii38Dec. 1998Hosking & MahrRobertson, NSW−33.5925, 150.6089AD99851269
EB19R. leucostachys 6Dec. 1997Bruzzese, Mahr & EvansCobden to Port Campbell Road, Vic.−38.5426, 143.0173AD99811454/5
EB 9R. leucostachys 7Dec. 1997Bruzzese, Mahr & Evansbetween Tallangatta and Corryong, Vic.−36.1779, 147.5474AD99811450/1
EB16R. leucostachys 7Dec. 1997Bruzzese, Mahr & EvansCobden, Vic.−38.3398, 143.0512AD989811458
960804R. leucostachysClose to 7Jul. 1996MahrFrankston, Vic.−38.1149, 145.1701Unavailable
971901R. leucostachys 7Dec. 1996MahrMurrongowan, Vic.−37.6319, 148.7239Unavailable
JH1669R. leucostachys 9Dec. 1998Hosking & Mahrnear Gibbo Arm of Dartmouth Dam, Vic.−36.7803, 147.7111AD99851323
972104R. leucostachys17Apr. 1997MahrFoster North, Vic.−38.1149, 145.1701Unavailable
972101R. leucostachys21Apr. 1997MahrFoster North, Vic.−38.1149, 145.1701Unavailable
15734R. phaeocarpus19Feb. 1998SymonLenswood, SA−34.9167, 138.8333AD99801028/9
961107R. polyanthemus36Aug. 1996MahrCallignee, Vic.−38.3582, 146.5705Unavailable
SR10 (15710)R. rubritinctus18Dec. 1996SymonCharleston to Mt Torrens Road, SA−34.9667, 138.9AD99750230
EB18R. sp. 2Dec. 1997Bruzzese, Mahr & EvansCobden to Port Campbell Road, Vic.−38.5426, 143.0173AD99811456
971702R. sp. 2Dec. 1996MahrBenambra, Vic.−36.9772, 147.7178Unavailable
15877R. sp. 3Dec. 1998Symon, Hosking & Mahrnorth of Ballarat, Vic.−37.5063, 143.8284AD99851300
SR43 (15711A)R. sp.14Dec. 1997SymonPenwortham, SA−33.9167, 138.65AD99750231
15382R. sp.15Dec. 1996SymonWaterfall Gully, SA−34.9667, 138.6667AD996542148/9
971606R. sp.39Dec. 1996MahrBuffalo River, Vic.−36.7521, 146.6587Unavailable
981902R. sp. Tasmania? (J.R. Hosking 1551)16Jan. 1998Roush & FrodshamRosebery, Tas.−41.7818, 145.5355Unavailable
WP11R. ulmifolius10Dec. 1998Mahrsouth of Gingkin, NSW−33.8976, 149.9239Unavailable
SR36 (15421)R. ulmifolius30Dec. 1996SymonMt Osmond, SA−34.9667, 138.6667AD99652164
982002R. ulmifolius40Jun. 1998Dellowbetween Gingkin and Tuglow, NSW−33.8889, 149.9281Unavailable
EB21R. vestitus28Dec. 1997Bruzzese, Mahr & EvansGellibrand River, Vic.−38.7279, 143.2461AD99811443/4

Experimental design

In early spring of 2001, crowns of each Rubus clone listed in Table 1 were transferred to 30-cm-diameter pots containing the same potting mixture as described above. These potted plants were placed in a field at the Waite Campus, located in the Adelaide plains, 25 days before each plant was inoculated with P. violaceum, as described below. Rubus clones were arranged in a randomized block design with five replications, each block comprising plants arranged in a staggered row so that plants were at least 1 m apart. Rows were orientated perpendicular to the direction of the prevailing wind and approximately 20 m apart so that they received an even distribution of overhead irrigation. Overhead irrigation was applied at night for 3 min at 20·00, 00·00 and 02·00 h, respectively, three or four times per week. Plants were watered manually to supplement irrigation when conditions were hot and dry. Temperature and rainfall data were collected from a weather station located approximately 500 m from the field. Degree days were calculated daily by subtracting a base temperature of 6°C (Jennings, 1988) from the daily average air temperature. Rainfall was noted because the airborne concentration of fungal conidia, in general, is thought to be correlated negatively with rain and positively with disease incidence (Blanco et al., 2004).

Pathogen isolate and inoculation

All plants in the randomized blocks were inoculated with urediniospores of P. violaceum isolate SA1, collected in the Adelaide Hills of South Australia in 1998 and in a region adjacent to the site of this study (Evans et al., 2000). Isolate SA1 of P. violaceum originated from a single uredinium, and was multiplied and stored as described by Evans et al. (2000). Voucher specimens of uredinia of isolate SA1 on (inoculated) leaflets of R. vestitus clone EB21, R. leucostachys clone JH1669 and R. sp. clone 971606 are stored at the Plant Pathology Herbarium of the New South Wales Department of Primary Industries as DAR 75508, DAR 75509 and DAR 75561, respectively. It is possible that isolate SA1 is related closely to the original unauthorized population of P. violaceum released in Victoria on or before 1984 (Marks et al., 1984). Isolate SA1 and 14 other single-uredinium isolates of P. violaceum collected in Australia recently did not share any HaeIII restriction fragments of DNA with P. violaceum isolate F15 (Evans et al., 2000), the authorized biological control agent, suggesting that these isolates were descendants of the original unauthorized population.

The most vigorous shoot per plant was selected for inoculation when new shoots had a median length of 12·5 cm. Shoot lengths ranged from 1 to 46 cm following variable rates of bud burst. A suspension of 0·25 mg mL−1 of urediniospores in water was applied as a fine mist with a hand-held atomizer to the abaxial surface of all leaves on the shoot selected. Plants were inoculated in the late afternoon of 9 October 2001, and then covered with plastic bags to ensure high humidity for germination and infection by P. violaceum. Plastic bags were removed the following morning at least 16 h after inoculation. On the same day, 20 plants from the same shade-house batch, representing Rubus clones of taxa EB21 and 9607, were transferred to the field site as noninoculated controls. Four plants were spaced evenly along each of the five inoculated blocks of plants and 1 m from the closest inoculated plant to check qualitatively for natural infection or secondary spread of the disease. The experimental site was located approximately 5 km from the nearest naturalized blackberry infestation and inoculation was timed for the first infection period to occur as early as possible in the spring. Disease symptoms developed on uninoculated plants as the second generation of uredinia appeared on inoculated plants, an observation that was consistent with spread of disease from inoculated plants as opposed to arrival of immigrant spore inoculum or pre-existing infection.

In addition to maintaining high humidity throughout evenings when overhead irrigation was supplied, disease development was promoted by providing a source of airborne urediniospores of P. violaceum isolate SA1. Ten potted plants of Rubus clone 9607 with erumpent uredinia were located randomly across the trial area 12 days after the test plants had been inoculated. These plants were inoculated with P. violaceum isolate SA1 as described previously. Inoculated plants were incubated in a tent and maintained with a fine coverage of water mist for 24 h. The plants were then moved to a glasshouse bench and uredinia began to appear 5 days after inoculation. These diseased plants were then moved to the field 16 days after they had been inoculated.

Assessment of shoot growth, disease and data analyses

To estimate the rate of leaf emergence in plants that had been inoculated with P. violaceum, the first fully expanded leaf on each inoculated shoot was marked with a string tag at 0, 8, 15, 22, 28 and 44 days postinoculation. The number of leaves between each tag on the shoot was counted and the cumulative number of leaves appearing from day 0 to each time of leaf tagging was calculated. Using simple linear regression, the relationship between the cumulative number of leaves appearing and time expressed as cumulative degree-days was estimated. The slope of the linear model was the number of leaves emerging per degree-day. Shoot length was also measured 61 days after inoculation.

The latent period of disease for inoculated plants was estimated as the cumulative degree-days from the time of inoculation to when approximately 50% of uredinia had erupted. The number of leaves emerging per latent period of uredinia (LELPU) was estimated as the latent period (cumulative degree-days) multiplied by the number of leaves emerging per degree-day.

Uredinia and telia were counted for disease assessment as well as estimating the percentage area of leaf necrosis. Counting of sori was done 24 and 58 days postinoculation, after the first generation of uredinia and after an estimated four generations of uredinia, respectively. Total numbers of uredinia, and telia if present, were counted manually on all compound leaves of a vegetative shoot that had at least one leaflet expressing signs of disease. The total number of uredinia and telia on each leaflet was allocated to one of four number categories: 0, 1–9, 10–100 or greater than 100 rust pustules. At 62 days postinoculation, disease severity was estimated as the percentage area of each leaflet that was necrotic for all leaves older than the first fully expanded leaf tagged 44 days postinoculation. Severely diseased leaflets that had fallen from the shoot were given a maximum score of 100 and the number of replicates with defoliation caused by rust disease was noted for each Rubus clone. The sum of percentage necrotic area for all leaflets assessed was calculated and divided by the maximum possible disease, calculated as the number of leaflets assessed multiplied by 100. For convenience, this assessment will be referred to as shoot disease severity. Disease severity was also estimated as the mean percentage necrotic leaflet area for all leaflets of three compound leaves that represented a zone of the shoot expressing the greatest amount of disease. This assessment was made to quantify disease on leaves that had maximum susceptibility at the time of infection, assuming the presence of leaf age-related disease resistance. The number of replicates where telia were produced was also noted for each Rubus clone. Data were analysed by one-way analysis of variance (anova, randomized complete blocks) using GenStat, release 6·1 (VSN International Ltd). If assumptions underlying anova were not satisfied, even after logarithmic transformation of continuous data or arcsine transformation of percentage data, then Friedman's (nonparametric) test was applied using the same version of GenStat.

Macroscopically, a Rubus clone was rated as being highly resistant to disease caused by P. violaceum isolate SA1 when no uredinia developed in any replicate of the trial at 24 or 58 days after inoculation. In addition, a Rubus clone was rated resistant to rust disease if uredinia or telia were produced at 58 days after inoculation but the mean percentage of leaflets with greater than nine pustules was less than 5% and there was no defoliation caused by rust disease. Otherwise, a Rubus clone was rated susceptible to rust disease without further categorization.

The relationship between mean disease severity of susceptible Rubus clones and mean LELPU was analysed by simple linear regression and by calculating Spearman's rank correlation coefficient (rs; Siegel, 1956). Four measures of mean disease severity were analysed separately: (i) mean number of uredinia, and telia if present, for diseased leaves only at 58 days after inoculation (dai); (ii) mean percentage of leaflets with greater than nine pustules for diseased leaves only at 58 dai; (iii) mean disease severity of the shoot at 62 dai; and (iv) mean disease severity for the three leaves expressing the greatest symptoms at 62 dai. The significance of rs, with a null hypothesis that mean disease severity and mean LELPU are not associated, was tested by the t-distribution on n − 2 degrees of freedom (d.f.).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The first symptoms of rust disease were observed on susceptible Rubus clones 12 days after inoculation. The latent period of the first generation of uredinia was estimated to be 16 days after inoculation or 150 degree-days, and urediniospores were shed visibly from uredinia 20 days after inoculation. The second generation of uredinia was observed 26 days after inoculation. Daily average temperature ranged from 11·3 to 26·6°C with a mean of 16°C (Fig. 1). The daily maximum temperature ranged from 15·1 to 31·5°C and the daily minimum temperature ranged from 6·4 to 18·4°C. There were five rainfall events that exceeded 5 mm (Fig. 1).


Figure 1. Daily average temperature (°C, line graph) and rainfall (mm, bar graph) at the Waite Campus, University of Adelaide, for 62 days after inoculation of test plants with Phragmidium violaceum in 2001.

Download figure to PowerPoint

There were significant differences (P < 0·001) among Rubus clones in their rate of leaf emergence during the first 44 days after inoculation (Table 2). The mean LELPU ranged from 5·2 for R. leightonii clone JH1660 to 8·7 for R. ulmifolius clone 982002. There were also significant differences in the rate of leaf emergence among different clones of the same species. For example, R. anglocandicans clone 9607 had a slower rate of leaf emergence when compared with R. anglocandicans clone 960204. Shoot lengths 61 days after inoculation ranged from 93 to 280 cm for all plants of the R. fruticosus agg. The final shoot length of the North American blackberry, R. laudatus clone JH 1656, ranged from 79 to 89 cm, and it appeared to grow at a rate similar to the slower-growing clones of the R. fruticosus agg.

Table 2.  Rate of leaf emergence and shoot growth of Rubus clones inoculated with Phragmidium violaceum isolate SA1. The number of leaves emerging per latent period of uredinia (LELPU) was estimated as the latent period of uredinia (cumulative degree-days) multiplied by the number of leaves emerging per degree-day
Rubus taxonRubus cloneMean LELPU during 44 days after inoculationRange in shoot length (cm), 61 days after inoculation
R. ulmifolius9820028.71154–188
R. vestitusEB218.71196–204
R. erythropsEB207.98171–272
R. polyanthemus9611077.77136–241
R. sp.153827.74123–212
R. sp.9716067.63142–213
R. anglocandicans9602047.62213–278
R. ulmifoliusWP117.41111–191
R. sp. Tasmania?9819027.34192–234
R. echinatus01017.28163–213
R. rubritinctusSR107.26159–184
R. leucostachys9721047.23158–253
R. leucostachys9721016.95156–219
R. leucostachysEB166.86132–150
R. cissburiensisEB146.85212–280
R. ulmifoliusSR366.78120–163
R. sp.158776.63140–186
R. sp.EB186.58136–175
R. leightoniiJH15666.48219–251
R. leucostachys9608046.30127–237
R. laciniatusKE16.21165–207
R. phaeocarpus157346.13222–250
R. sp.SR436.12113–237
R. leucostachys9719016.09155–170
R. anglocandicans96075.97109–225
R. leucostachysEB195.90144–161
R. sp.9717025.86129–171
R. laudatusJH16565.83 79–89
R. leucostachysEB95.75153–218
R. laciniatusSR145.33127–141
R. leucostachysJH16695.29103–172
R. leightoniiJH16605.21 93–178
s.e.d. 0.61 (d.f. = 114)

Variation in susceptibility of Rubus clones to isolate SA1 was evident after production of the first generation of uredinia (Table 3). Assessment of uredinia 24 days after inoculation revealed that R. anglocandicans clones 9607 and 960204, and R. leightonii clone JH1566 were highly susceptible with greater than 20% of leaflets assessed with more than 100 uredinia. No uredinia were produced on leaflets of R. sp. clone 971702, R. cissburiensis clone EB14 and R. laudatus clone JH1656. Rubus sp. clone SR43 and R. leucostachys clones EB19, EB9 and 960804 had no leaflets with numbers of uredinia in the categories 10–100 or > 100. Other clones of R. leucostachys (EB16, 971901, JH1669, 972104 and 972101) had 0–9% of leaflets assessed with uredinia in the > 100 number category, and clone 971901 had a very low mean number of uredinia per leaflet. Rubus erythrops clone EB20 also had no leaflets with uredinia in the > 100 number category and a low mean number of uredinia per leaflet. When compared with other Rubus clones, the two clones of R. laciniatus (SR14, KE1) produced an intermediate and similar response to isolate SA1, as did the three clones of R. ulmifolius (982002, WP11 and SR36). Other susceptible clones were R. echinatus clone 0101, R. phaeocarpus clone 15734, R. polyanthemus clone 961107, R. rubritinctus clone SR10, R. sp. clones EB18, 15877, 15382 and 9971606, R. sp. Tasmania? clone 981902 and R. vestitus clone EB21.

Table 3.  Rating of Rubus clones for susceptibility to disease caused by Phragmidium violaceum isolate SA1 based on numbers of uredinia and telia on leaflets of all compound leaves that had at least one leaflet with rust pustules, 24 and 58 days after inoculation (dai). Each leaflet was allocated to a number category (0, 1–9, 10–100 or > 100) if the number of pustules present on the leaflet matched the range indicated by the number category
Rubus taxonRubus cloneDisease RatingaMean number of uredinia, and telia if present, per leafletbMean percentage of leaflets with uredinia, and telia if present, in each number category
24 days after inoculationc58 days after inoculation
24 daic58 dai01–910–100> 10001–910–100> 100
  • a

    R, resistant; S, susceptible to disease caused by P. violaceum isolate SA1 (refer to Table 4 for number of replicates with leaf drop).

  • b

    Clones sorted on mean number of pustules per leaflet, at 24 dai. Leaflets that had greater than 100 pustules were given a maximum score of 100.

  • c

    Data from four replicates (blocks) only.

  • d

    n= 3 and df = 28 as plots with missing data were not included in the analysis.

  • e

    n= 5 and df = 30 as R. echinatus clone 0101 was excluded from the analysis.

  • f

    Dashes represent data not suitable for analysis of variance.

  • g

    Data for R. sp. clone 971702, R. cissburiensis clone EB14 and R. laudatus clone JH1656 were excluded from the analysis of variance.

  • h

    Data for Rubus clones 971901, EB20, EB19, EB9, 960804, SR43, 971702, EB14 and JH1656 were excluded from the analysis of variance.

R. anglocandicans960204S6722  3·3 3·944·947·8 14·940·836·0 8·4
R. anglocandicans9607S5844  2·3 6·861·629·4  5·522·144·627·8
R. leightoniiJH1566S4920  7·514·255·722·6 10·441·943·8 3·9
R. sp.15382S4151  5·017·573·8 3·8  5·516·150·128·3
R. vestitusEB21S4117  9·222·557·510·8 12·746·536·2 4·4
R. sp.EB18S4030  5·819·463·611·1  4·428·260·3 7·2
R. echinatus0101S3932 17·62948·4 4·9  9·825·359·4 5·6
R. leucostachys972101S3832 18·6 9·363·5 8·6  2·626·762·5 8·2
R. laciniatusSR14S3832  7·634·152·6 5·8  5·437·746·110·8
R. laciniatusKE1S3625  9·714·671·5 4·2  5·733·252·7 8·3
R. sp.971606S2822 22·219·454·9 3·5 13·542·837·5 6·2
R. sp. Tasmania?981902S2820 10·533·643·812·1 15·738·440·8 5·1
R. leucostachysEB16S2711  6·223·866·9 3·1 11·257·231·1 0·5
R. ulmifolius982002S25 9 11·731·253·3 3·8 15·362·219·9 2·6
R. ulmifoliusWP11S2412 12·229·352·5 6·0 14·748·536·3 0·5
R. leightoniiJH1660S2418  2·840·356·9 0 13·944·836·3 5·0
R. polyanthemus961107S2320 10·618·571 0 11·636·051·2 1·1
R. leucostachys972104S2223 14·425·846·9 4·2  7·338·450·3 4·0
R. leucostachysJH1669S2212 33·825·336·6 4·32  9·036·852·9 1·2
R. phaeocarpus15734S2231 19·435·142·6 2·8 10·830·445·713·1
R. rubritinctusSR10S2113 15·027·653·7 3·7 12·942·544·6 0·0
R. sp.15877S1614 12·738·944·6 3·8 14·849·134·4 1·7
R. ulmifoliusSR36S1616 14·443·838·1 3·7 11·050·833·8 3·4
R. leucostachys971901R 3 0·1 77·5 9·213·3 0 98·2 1·8 0 0
R. erythropsEB20R 2 0·06 67·425·7 6·9 0 95·5 4·5 0 0
R. leucostachysEB19R 0·2 0·01 88·711·3 0 0 99·3 0·7 0 0
R. leucostachysEB9R 0·3 0 89·610·9 0 0100 0 0 0
R. leucostachys960804R 0·04 0·3 95·8 4·2 0 0 97·5 0·8 1·7 0
R. sp.SR43R 0·2 0 91·4 8·6 0 0100 0 0 0
R. cissburiensisEB14R 0 0·02100 0 0 0 98·4 1·6 0 0
R. sp.971702R 0 0100 0 0 0100 0 0 0
R. laudatusJH1656R 0 0100 0 0 0100 0 0 0
P-value for Friedman's test s.e.d.f   0·000d 0·000e  0·001d 0·03d 0·003d 0·05d  0·000e 0·000e 0·000e 0·000e
  13·8g (d.f. = 80) 9·1h (d.f. = 85)

After approximately four generations of uredinia, the response of Rubus clones showed the same trend in their susceptibility to isolate SA1 (Table 3). By this time, 58 days postinoculation, telia had formed in some or all replicates of various Rubus clones. The only Rubus clones with greater than 20% of leaflets assessed with > 100 uredinia and telia were R. anglocandicans clone 9607 and R. sp. clone 15382. Again, no pustules were produced on leaflets of R. sp. clone 971702 and R. laudatus JH1656. Unlike the previous assessment, no pustules were recorded for Rubus sp. clone SR43 and R. leucostachys clone EB9, and very low numbers of pustules were recorded for R. cissburiensis clone EB14. As before, R. leucostachys clone EB19 had no leaflets with greater than nine sori, although R. leucostachys clone 960804 had 1·7% of leaflets with numbers of pustules in the 10–100 category. Rubus erythrops clone EB20 and R. leucostachys clone 971901 remained resistant with < 10 sori per leaflet. All remaining host genotypes were susceptible with numbers of pustules on leaflets ranging over the 10–100 category depending on genotype.

Mean disease severity of the shoot for all Rubus clones assessed 62 days after inoculation ranged from 0·3 to 38% (Table 4). Rubus clones identified as highly susceptible by pustule counts had mean shoot disease severities of 18, 29, 37 and 38% for R. leightonii clone JH1566, R. anglocandicans clones 960204 and 9607 and R. sp. 15382, respectively. These taxa also had a high mean disease severity for the three leaves most rust affected (71, 98, 91 and 92%, respectively) together with R. vestitus clone EB21 and R. laciniatus clone SR14.

Table 4.  Mean disease severity estimated as area of necrotic leaf tissue, incidence of leaf drop and telia on Rubus clones 62 days after inoculation with Phragmidium violaceum isolate SA1
Rubus taxonRubus cloneMean disease severitya of shootMean disease severitya for the three leaves exhibiting greatest disease symptomsNumber of replicates with leaf drop caused by rust diseaseNumber of replicates producing telia
  • a

    Disease severity calculated as the sum of necrotic leaflet area as a percentage of total leaflet area assessed.

  • b

    Four replicates assessed.

  • c

    Discrete data were not subjected to anova.

R. sp.15382 37.8 92.052
R. anglocandicans9607 37.2 91.452
R. laciniatusSR14 36.9 95.040
R. vestitusEB21 33.410033
R. anglocandicans960204 29.4 98.153
R. laciniatusKE1 26.7 86.642
R. ulmifoliusWP11 23.9 73.830
R. ulmifolius982002 20.6 84.943
R. sp. Tasmania?981902 18.7 93.045
R. sp.971606 18.4 80.522
R. leucostachys972104 18.3 62.622
R. leightoniiJH1566 17.6 70.921
R. leightoniiJH1660 17.0 50.814
R. sp.EB18 15.6 53.034
R. ulmifoliusSR36 14.4 60.441
R. leucostachys972101 13.9 49.613
R. phaeocarpus15734 13.5 45.741
R. sp.15877 10.9 36.023
R. echinatus0101 10.9 35.50b1b
R. rubritinctusSR10 10.3 44.31b1b
R. leucostachysJH1669  9.0 21.710
R. polyanthemus961107  8.3 44.012
R. sp.971702  5.5 21.000
R. leucostachysEB16  4.7 14.524
R. leucostachys971901  2.0  3.700
R. sp.SR43  1.2  3.000
R. laudatusJH1656  1.1  5.100
R. leucostachysEB9  1.1  1.700
R. leucostachysEB19  1.1  2.800
R. erythropsEB20  1.1  4.500
R. leucostachys960804  0.9  3.300
R. cissburiensisEB14  0.3  0.600
s.e.d.   4.42 11.2cc
d.f. 110114

The pattern of disease across a whole shoot for two susceptible clones, 15382 and 971606, is illustrated in Table 5. In both cases, pustule numbers appeared to rise and fall at least twice with leaf position. For the 23 Rubus clones rated as susceptible to rust disease (Table 3), simple regression analyses of four different measures of disease severity and mean LELPU were insignificant at P = 0·05. However, two measures of mean disease severity were associated weakly with mean LELPU when Spearman's rank correlation coefficient was calculated (Fig. 2). These measures were mean percentage of leaflets with greater than nine sori (diseased leaves only) and mean disease severity for the three leaves expressing greatest disease symptoms, measured at 58 and 62 days after inoculation, respectively. The association was negative (rs=−0·28) for the former and positive (rs= 0·28) for the latter measurement of disease severity (0·10 < P > 0·05).

Table 5.  Example of the distribution of rust pustules on leaves of a single shoot for two susceptible Rubus clones, 58 days after inoculation with Phragmidium violaceum isolate SA1
Leaf positionaNumber of uredinia, and telia if present, per leaflet
Rubus clone 15382, replicate 5Rubus clone 971606, replicate 3
Left lateralTerminalRight lateralLeft lateralTerminalRight lateral
  • a

    Leaf position is relative to the first fully expanded leaf (leaf position 1) on the day of inoculation. Leaf 3 is the oldest leaf relative to the others listed.

  • b

    lm, leaflet missing following severe disease.

  • c

    The dash indicates that the shoot had not yet developed a leaf at this position.

386> 100> 100222629
4lmblm> 100376342
5> 100> 100> 100615143
6> 100> 100> 10066> 10093
7> 100> 100> 100647732
8> 100> 100> 100466443
10> 100> 10081224627
1173938572> 100> 100
12> 100> 100> 100363941
1389> 100> 100272628
147134> 100 410 2
15461741 0 4 0
16523928 2 2 2
17 9 2 6 5 4 1
18 6 811 3 1 3
19 1 0 2 2 813
20 0 1 0 6 6 1
21 1 0 01623 5
22 0 0 0 0 4 3
23 0 0 0 0 6 1
24 0 0 0 1 0 0
25 0 0 0 0 0 0
26 0 0 0 0 0 0
27 0 0 0 0 0 0
28 0 0 0 0 0 0
29 0 0 0 0 0 0
30c 0 0 0
31 0 0 0
32 0 0 0

Figure 2. The association between two measures of mean disease severity and mean leaves emerging per latent period of uredinia (LELPU) in the range 5–9 for susceptible Rubus clones (n = 23, d.f. = 21). Mean disease severity was measured as mean percentage of leaflets with greater than nine pustules of the rust, Phragmidium violaceum (diseased leaves only, two number categories combined from Table 3) (a), or mean disease severity for the three leaves expressing greatest disease symptoms (b).

Download figure to PowerPoint

In summary, clones exhibiting high resistance to rust disease (no rust pustules produced) were R. sp. clone 971702 and R. laudatus clone JH1656. Resistant phenotypes with a few pustules were R. sp. clone SR43, R. leucostachys clones 971901, EB9, EB19 and 960804, R. erythrops clone EB20, and R. cissburiensis clone EB14. Two of these, R. leucostachys clones EB9 and 971901, share the same DNA phenotype with R. leucostachys clone EB16, which was rated susceptible. Variation in mean LELPU among the 23 susceptible Rubus clones probably accounted for some of the variation observed in mean disease severity, although the correlation was weak and not significant (P ≤ 0·05).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

A broad sample of taxa of the R. fruticosus agg., isolated from naturalized infestations in Australia, varied in susceptibility to rust disease caused by the uredinial stage of a single isolate of P. violaceum (SA1) collected recently in Australia. Furthermore, clones of R. leucostachys with the same DNA phenotype expressed different responses to infection by isolate SA1. Bruzzese & Hasan (1986) reported that P. violaceum (F15) isolated in Europe, unlike isolate SA1 used here, did not produce uredinia following controlled-environment inoculation and incubation of detached leaves of R. laciniatus, a distinct and readily identifiable blackberry taxon. These combined data provide evidence of variation in pathogenicity among isolates of P. violaceum as in other species of Phragmidium and their hosts (e.g. Anthony et al., 1985).

Preliminary studies of genetic diversity of P. violaceum in Australia between 1997 and 1999 indicate that this population was genetically variable but with a low diversity typical of that associated with a founder effect (Evans et al., 2000). That eight out of 31 clones of the R. fruticosus agg. were found to exhibit resistance to rust disease caused by isolate SA1, isolated in Australia in 1998, might explain the absence of severe rust disease in some blackberry infestations.

The response of blackberry clones to infection by other spore stages of P. violaceum is unknown. In Australia, teliospores mature on leaves that are not shed from primocanes during winter. In spring, germinating teliospores release basidiospores which give rise to spermagonia and aecia mainly on floricane leaves. In a comprehensive study of the inheritance of resistance in red raspberry to yellow rust, Anthony et al. (1986) assessed incidence and severity of infection at the telial stage of P. rubi-idaei, in addition to numbers of uredinia and latent period of uredinia. This approach allowed investigation of components of resistance throughout the growing season and the telial index (exposed bait plants) and uredinial (inoculated) stages appeared to be positively correlated. Red raspberry Malling Jewel expressed some resistance to P. rubi-idaei at the telial stage but developed more aecia per leaf and greater incidence at the aecial stage when compared with two other red raspberry cultivars. Similarly, investigating host resistance at all stages in the life cycle of P. violaceum may be informative.

Both clones of R. anglocandicans were susceptible to isolate SA1 of P. violaceum, producing relatively high numbers of uredinia per leaflet. Both of these clones had significantly different rates of leaf emergence in the presence of rust disease, perhaps reflecting differences in plant origin (unknown) or response to the different environments during naturalization. This taxon of the R. fruticosus agg., named previously as R. procerus or R. discolor, is widespread throughout Australia and appears to exist as a single clonal lineage (Evans & Weber, 2003). It is also a common weedy blackberry in the Adelaide Hills, suggesting that suitable weather conditions should produce severe rust disease on this Rubus species. Severe rust disease on R. anglocandicans growing in a humid microclimate has been observed in this location. Such rust epidemics have been halted in early summer when telia formed after a succession of days when the temperature exceeded 30°C and during summer drought when active growth of blackberry shoots ceased. Like R. anglocandicans, two clones of R. laciniatus were assayed and found to have a similar level of susceptibility to rust disease. Both clones of R. laciniatus have the same DNA phenotype and are genetically similar, but their leaf morphology differs. Clone SR14 has leaflets that are deeply divided, whereas clone KE1 has leaflets that are undivided. This difference in leaf morphology does not appear to affect susceptibility to rust disease.

Unlike R. anglocandicans, R. leucostachys is a genetically and morphologically variable taxon of the R. fruticosus agg., with at least seven DNA phenotypes (Evans et al., 2004b). At least two DNA phenotypes are resistant to infection by P. violaceum (SA1), thus demonstrating phenotypic variation in this Rubus species as well. Conversely, all three clones of R. ulmifolius, representing three different DNA phenotypes of this diploid sexual species, were similar in their response to infection by isolate SA1. Rubus laudatus clone JH1656 is a blackberry of North American origin that has no history of coevolution with P. violaceum. The lack of uredinia forming on this Rubus clone was consistent with its origin, even though it scored a mean of 6% for disease severity based on necrotic symptoms on three most affected leaves. These symptoms, however, may have been due to a hypersensitive response to infection, infection and colonization without sporulation or an unrelated biotic/abiotic event.

Disease severity was estimated under field conditions conducive to development of disease and host growth. While maximum mean disease severity of a single shoot did not exceed 40%, disease was severe enough to cause leaf drop when mean disease severity of the shoot was as low as 5% (e.g. R. leucostachys clone EB16). In controlled-environment conditions, Evans & Bruzzese (2003) demonstrated that leaf age explained 97% of the variation in uredinium production by P. violaceum (F7) on leaflets of shoots of R. anglocandicans growing at different rates on a single plant. Therefore, leaf drop is likely to occur when an individual leaf is highly susceptible at the time of infection. A constant source of numerous, viable airborne urediniospores would ensure uniform exposure of leaves to inoculum as they develop maximum susceptibility to infection. However, the concentration of airborne fungal spores is known to vary depending on key environmental variables, the strength of the sporulating source and distance from it (Lacey, 1996). The fact that all leaves along the shoot did not have constant sorus numbers, whereby sorus numbers rose and fell with leaf position, suggests that leaves were emerging faster than the latent period for urediniospore production. Indeed, there were five to nine leaves produced per estimated latent period. Furthermore, there was a weak and negative association between mean LELPU and mean percentage of leaflets with greater than nine sori 58 days after inoculation. A possible explanation for this negative association is that leaves emerged rapidly then aged and became more resistant to infection before exposure to the next temporal peak in airborne urediniospores. There was also a weak and positive association between mean LELPU and mean disease severity for the three most severely affected leaves. Presumably Rubus clones with mean LELPU at the higher end of the range tended to have leaves at adjacent nodes that were close in age and highly susceptible at a time of infection when the airborne concentration of urediniospores was high. The lack of a clear association between mean disease severity and mean LELPU suggests that a wider range of LELPU should be generated in future experiments, both within and among Rubus clones. Additional isolates of P. violaceum should also be tested to explore fully the impact of age-related disease resistance under field conditions.

While there is likely to be an optimum rate of host growth and pathogen reproduction that results in effective biocontrol, disease levels in the field might also be a function of shoot density. In particular, the higher the shoot density, the larger the number of leaves per canopy area that are in the appropriate age category for significant infection by P. violaceum. There might also be an increase in relative humidity within a dense blackberry canopy, which would be conducive to development of rust disease. To illustrate variation in shoot density, Amor (1975) reported that the density of live canes of R. anglocandicans averaged 18 canes m−2, in contrast to a blackberry named by the same author as ‘R. ulmifolius hybrid’ which averaged 52 canes m−2. It is likely that Amor's ‘R. ulmifolius hybrid’ was a biotype of R. leucostachys (Evans et al., 2004b).

In conclusion, counts of uredinia and telia enabled categorization of Rubus clones as either susceptible or resistant to rust disease even though mean disease severity appeared continuous over the range of Rubus clones tested. DNA phenotyping of Rubus clones provided precise characterization of fine-scale variation in susceptibility to rust disease when two Rubus clones with the same DNA phenotype interacted differently with P. violaceum isolate SA1. The mechanism of resistance to rust disease in the field requires further study given that there was no clear association between mean LELPU and mean disease severity over the range in LELPU observed.

The results of this field study have been used for evaluating and validating a controlled-environment bioassay for identifying physiological specialization in P. violaceum and subsequent pathotyping of rust isolates deployed as biological control agents. The preliminary report of physiological specialization in P. violaceum can be found in Evans & Gomez (2004) and Evans et al. (2004a).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The Cooperative Research Centre for Australian Weed Management supported this work. Special thanks go to D. E. Symon (State Herbarium of South Australia), J. R. Hosking (NSW Department of Primary Industries) and E. Bruzzese and F. A. Mahr (Department of Primary Industries, Victoria) for collection of Rubus crowns and for leading the authors to relevant collection sites.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Alice LA, Campbell CS, 1999. Phylogeny of Rubus (Rosaceae) based on nuclear ribosomal DNA internal transcribed spacer region sequences. American Journal of Botany 86, 8197.
  • Alice LA, Eriksson T, Eriksen B, Campbell CS, 2001. Hybridisation and gene flow between distantly related species of Rubus (Rosaceae): evidence from nuclear ribosomal DNA internal transcribed spacer region sequences. Systematic Botany 26, 76978.
  • Amor RL, 1975. Ecology and control of blackberry (Rubus fruticosus L. agg.). IV. Effects of single and repeated applications of 2,4,5-T, picloram and aminotriazole. Weed Research 15, 3945.
  • Amor RL, Richardson RG, Pritchard GH, Bruzzese E, 1998. Rubus fruticosus L. Agg. In: PanettaFD, GrovesRH, ShepherdRCH, eds. The Biology of Australian Weeds, Vol. 2. Melbourne, Australia: RG & FJ Richardson, 22546.
  • Anthony VM, Shattock RC, Williamson B, 1985. Interaction of red raspberry cultivars with isolates of Phragmidium rubi-idaei. Plant Pathology 34, 5217.
  • Anthony VM, Williamson B, Jennings DL, Shattock RC, 1986. Inheritance of resistance to yellow rust (Phragmidium rubi-idaei) in red raspberry. Annals of Applied Biology 109, 36574.
  • Blanco C, De Santos B, Los Barrau C, Arroyo FT, Porras M, Romero F, 2004. Relationship among concentrations of Sphaerotheca macularis conidia in the air, environmental conditions, and the incidence of powdery mildew in strawberry. Plant Disease 88, 87881.
  • Bruzzese E, Hasan S, 1986. Collection and selection in Europe of isolates of Phragmidium violaceum (Uredinales) pathogenic to species of European blackberry naturalized in Australia. Annals of Applied Biology 108, 52733.
  • Evans KJ, Bruzzese E, 2003. Life history of Phragmidium violaceum in relation to its effectiveness as a biological control agent of European blackberry. Australasian Plant Pathology 32, 2319.
  • Evans KJ, Gomez DR, 2004. Genetic markers in rust fungi and their application to weed biocontrol. In: EhlerLE, SforzaR, MateilleT, eds. Genetics, Evolution and Biological Control. Wallingford, UK: CABI Publishing, 7396.
  • Evans KJ, Jones MK, Mahr FA, Roush RT, 2000. DNA phenotypes of the blackberry biological control agent, Phragmidium violaceum, in Australia. Australasian Plant Pathology 29, 24954.
  • Evans KJ, Morin L, Bruzzese E, Roush RT, 2004a. Overcoming limits on rust epidemics in Australian infestations of European blackberry. In: CullenJM, BrieseDT, KriticosDJ, LonsdaleWM, MorinL, ScottJK, eds. Proceedings of the Eleventh International Symposium on Biological Control of Weeds, 2003. Canberra, Australia: CSIRO Entomology, 5149.
  • Evans KJ, Symon DE, Roush RT, 1998. Taxonomy and genotypes of the Rubus fruticosus L. aggregate in Australia. Plant Protection Quarterly 13, 1526.
  • Evans KJ, Symon DE, Whalen MA, Barker RM, Hosking JR, Oliver JA, 2004b. Taxonomic update and Lucid key for introduced blackberry in Australia. In: SindelBM, JohnsonSB, eds. Proceedings of the Fourteenth Australian Weeds Conference, 2004. Sydney, Australia: Weed Society of New South Wales, 50912.
  • Evans KJ, Weber HE, 2003. The most widespread taxon of weedy blackberry in Australia is Rubus anglocandicans (Rosaceae). Australian Systematic Botany 16, 52737.
  • Jennings DL, 1988. Raspberries and Blackberries: their Breeding, Diseases and Growth. London, UK: Academic Press Limited.
  • Lacey J, 1996. Spore dispersal – its role in ecology and disease: the British contribution to fungal aerobiology. Mycological Research 100, 64160.
  • Laundon GF, Rainbow AF, 1969. Phragmidium violaceum (Schultz) Wint. In: Descriptions of Pathogenic Fungi and Bacteria. Kew, UK: Commonwealth Mycological Institute. (No. 209.)
  • Mahr FA, Bruzzese E, 1998. The effect of Phragmidium violaceum (Shultz) Winter (Uredinales) on Rubus fruticosus L. agg in south-eastern Victoria. Plant Protection Quarterly 13, 1825.
  • Marks GC, Pascoe IG, Bruzzese E, 1984. First record of Phragmidium violaceum on blackberry in Victoria. Australasian Plant Pathology 13, 123.
  • Nybom H, 1988. Apomixis versus sexuality in blackberries (Rubus subgen. Rubus, Rosaceae). Plant Systematics and Evolution 160, 20718.
  • Pigott JP, Weiss J, Evans KJ, Mahr F, 2003. Predicting the effect of the rust disease (Phragmidium violaceum (Schultz) Winter) on blackberry (Rubus fruticosus L. agg.) in Victoria (Australia). In: BrunduG, BrockJ, CamardaI, ChildL, WadeM, eds. Plant Invasions: Species Ecology and Ecosystem Management. Leiden, The Netherlands: Backhuys Publishers, 37381.
  • Siegel S, 1956. Nonparametric Statistics for the Behavioral Sciences. New York, USA: McGraw-Hill.
  • Thompson JN, Burdon JJ, 1992. Gene-for-gene coevolution between plants and parasites. Nature 360, 1216.
  • Thorp JR, Lynch R, 2000. The Determination of Weeds of National Significance.Launceston, Australia: National Weeds Strategy Executive Committee.
  • Washington WS, 1985. The Susceptibility of Cultivated Rubus Varieties to Phragmidium violaceum, the Cause of Blackberry Leaf Rust, and Fungicides for the Control of the Disease. Adelaide, Australia: University of Adelaide, MSc Thesis.
  • Weber HE, 1996. Former and modern taxonomic treatment of the apomictic Rubus complex. Folia Geobotanica et Phytotaxonomica 31, 37380.