Biological Control 67 (2013) 350–360
Contents lists available at ScienceDirect
Biological Control journal homepage: www.elsevier.com/locate/ybcon
Successful establishment of epiphytotics of Puccinia punctiformis for biological control of Cirsium arvense Dana Berner a,⇑, Emily Smallwood a, Craig Cavin a, Anastasia Lagopodi b, Javid Kashefi c, Tamara Kolomiets d, Lyubov Pankratova d, Zhanna Mukhina e, Michael Cripps f, Graeme Bourdôt f a
USDA-ARS-Foreign Disease-Weed Science Research Unit, 1301 Ditto Avenue, Ft. Detrick, MD 21702-5023, USA School of Agriculture, Aristotle University of Thessaloniki, GR 54124 Thessaloniki, Greece USDA-ARS-European Biological Control Laboratory, Tsimiski 43, 7th Floor, GR 54623 Thessaloniki, Greece d All-Russia Research Phytopathology Institute, VNIIF, B. Vazemy, 143050 Moscow Region, Russia e All-Russia Rice Research Institute, Belozerny, 350921 Krasnodar, Russia f AgResearch Ltd., Lincoln, Private Bag 4749, Christchurch, New Zealand b c
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Rosettes of C. arvense were inoculated
in the fall with telia-bearing leaves. Rosettes inoculations were done in
Greece, New Zealand, Russia, and the USA. Conditions were favorable for teliospore germination in the fall in all fields. Systemically diseased shoots resulted from inoculations in all 13 field sites. Rosettes in the fall are the infection court for basidiospore infection.
a r t i c l e
i n f o
Article history: Received 2 April 2013 Accepted 11 September 2013 Available online 18 September 2013 Keywords: Biological control Canada thistle Cirsium arvense Epiphytotics Rust fungus Systemic disease
a b s t r a c t Canada thistle (Cirsium arvense, CT) is one of the worst weeds in temperate areas of the world. The rust fungus Puccinia punctiformis was first proposed as a biological control agent for CT in 1893. The rust causes systemic disease, is specific to CT, and is in all countries where CT is found. Despite a 120-year lapse since biological control with the rust was proposed, establishment of epiphytotics of the rust have previously been unsuccessful due to incomplete understanding of the disease cycle. In this study, newlyemerging rosettes in the fall are proposed as the physical and temporal infection courts for basidiospores, from germinating teliospores, to systemically infect CT and give rise to systemically diseased shoots the following spring. To test this hypothesis, rosettes of CT were inoculated in the fall with either telia-bearing leaves collected in mid-summer or with greenhouse-produced teliospores. Field sites were located near Kozani, Greece, Moscow, Russia, Christchurch, New Zealand, and Ft. Detrick, Maryland, USA. Teliabearing leaves, which were used as inoculum in 12 of 13 field sites, were collected near each field site from CT shoots in close proximity to systemically diseased CT shoots producing aeciospores in the spring. Aeciospore infections of the leaves of these nearby shoots gave rise to uredinia which turned to telia in mid- to late-summer. Temperature and dew conditions at inoculation in the fall at each site were very favorable for teliospore germination. Rosettes inoculated in the fall were marked with flags, and systemically diseased shoots emerging near these flags the following spring were recorded. In 11 of the sites in these countries, individual rosettes were inoculated 2, 4, 6, or 8 times with telia-bearing leaves. Proportions of rosettes giving rise to systemically diseased shoots, out of the number of rosettes inoculated, were analyzed. Inoculations in all 13 sites produced systemically diseased shoots. A separate study on the phenology of CT showed that the maximum rate of leaf abscission occurred at the time of maximum
⇑ Corresponding author. Fax: +1 301 619 2880. E-mail address:
[email protected] (D. Berner). 1049-9644/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.biocontrol.2013.09.010
D. Berner et al. / Biological Control 67 (2013) 350–360
351
emergence of new CT rosettes in the fall. This period coincided with an annually occurring period of sustained dew and favorable temperatures for teliospore germination. In nature, abscising telia-bearing leaves likely come into contact with a receptive rosette during favorable conditions for teliospore germination in the fall. This study demonstrates that epiphytotics of systemic rust disease of CT can be routinely established, by mimicking the natural disease cycle. Published by Elsevier Inc.
1. Introduction Canada thistle, also known as California thistle and creeping thistle, (Cirsium arvense (L.) Scop., Asteraceae, CT), is a perennial weed of pastures, rangelands and agricultural lands in the temperate areas of the world (Guiggisberg et al., 2012; Morishita, 1999). It is native to southeastern Europe and North Africa and is believed to have been introduced to North America from Europe in the 1600s as a contaminant in seed grain (Erickson, 1983; Guiggisberg et al., 2012). CT has now spread to 38 states in the U.S. and has been declared noxious in 25 (Kartesz and Meacham, 1999). It causes considerable, economically measurable, yield loss in small grains through both direct effects of competition (Donald, 1994) and indirect effects through allelopathy (Bendall, 1975; Stachon and Zimdahl, 1980). In pastures, parks, roadsides, and natural areas, control of CT often involves herbicides and mowing that are particularly costly on low-value land where these practices do not pay dividends in increased economic return (Amor and Harris, 1977; Beck and Sebastian, 2000; Trumble and Kok, 1982). Furthermore, none of these approaches are possible in pastures on steep hills such as those frequently infested by the weed in New Zealand and in parts of Greece and the USA. CT reproduces vegetatively by shoot buds on roots and sexually by seeds (Bakker, 1960; Donald, 1994), but sexual reproduction is not as important as asexual vegetative reproduction in thistle patch persistence and spread (Donald, 1994). Vegetative spread through horizontal growth of the root system is rapid especially when competition from other plants is low. As a result, CT typically grows in clonal patches from several meters up to 25 M or more in diameter (Donald, 1994). Plowing aggravates CT infestations by cutting up roots and producing root fragments which bud and produce new shoots (ramets) and new patches (genets) (Magnusson et al., 1987). Biological control using natural enemies offers a host-specific and low-cost management option for CT. Although several antagonistic organisms have been evaluated as biological control agents on CT, including fungi, bacteria, and insects (Bourdôt et al., 2006; Cripps et al., 2012; Green and Bailey, 2000; Gronwald et al., 2002), the obligate rust fungus Puccinia punctiformis (F. Strauss) Rohl. (=Puccinia sauveolens (Pers.) Rostr., =Puccinia obtegens (Link) Tul.) was perhaps the first plant pathogen proposed as a biological control agent for CT. In 1893 a New Jersey farmer noticed CT patches diseased with the rust virtually disappeared after a couple of years, and he proposed that the rust be widely disseminated to control the weed (Wilson, 1969). Puccinia punctiformis was probably introduced in N. America in rootstocks of CT (Olive, 1913) in the 1600s. The rust fungus is now naturalized in N. America and is found in all states and provinces in which CT is found (Farr and Rossman, 2013). It is very specific to CT (Guske et al., 2004). The fungus is an autoecious, brachy-form rust that can systemically infect CT (French and Lightfield, 1990). Systemic infections result in permanent infection of the roots (Cockayne, 1915; Menzies, 1953). Systemically diseased shoots arising from infected roots are relatively tall, spindly, frequently deformed (Fig. 1), and produce large amounts of aeciospores over an extended period of time (Van den Ende et al., 1987). Systemically diseased plants
Fig. 1. Shoot of Canada thistle systemically diseased with the rust fungus Puccinia punctiformis. The yellow color of the shoot is due to fragrant haploid spermagonia of the fungus. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
(genets) and all of the shoots (ramets) on infected root systems eventually die (Watson and Keogh, 1980). Newly-emerged systemically diseased shoots and leaves bear orange-colored haploid pycnia and receptive hyphae, which emit a sweet-smelling aroma similar to that of CT flowers (Rostrup, 1874; Connick and French, 1991). The fungus is heterothallic and requires fusion (plasmogamy) of opposite mating types of spermagonia (pycniospores and receptive hyphae) from different plants (Buller, 1950; Menzies, 1953). Following spermatization (plasmogamy) of the receptive hyphae by compatible spermagonia (pycniospores), the receptive hyphae give rise to brown aecia bearing dikaryotic aeciospores (Fig. 2), which become wind-born and cause local infections of neighboring thistle shoots and leaves later in the season (Cockayne, 1915; Olive, 1913; Rostrup, 1874; Waters, 1928). These local infections result in the development of localized lesions, bearing uredinia producing urediniospores on leaves of nearby shoots. The uredinia gradually become telia (Fig. 3) that produce teliospores in late summer (Cockayne, 1915; Bailiss and Wilson, 1967). Under favorable environmental conditions, teliospores germinate in response to an exogenous stimulant from the host plant and give rise to a basidium with four haploid basidiospores (Fig. 4) (Buller, 1950; French and Lightfield, 1990; Van den Ende et al., 1987). The basidiospores infect the host plant, and the rust overwinters as mycelium in the rootstock (Bailiss and Wilson, 1967; Olive, 1913; Rostrup, 1874).
352
D. Berner et al. / Biological Control 67 (2013) 350–360
Fig. 4. Two-celled diploid teliospores of Puccinia punctiformis germinating into a basidium and one-celled haploid basidiospores (upper center).
Fig. 2. Shoot of Canada thistle systemically diseased with the rust fungus Puccinia punctiformis. The undersides of the leaves are covered with dikaryotic aecia producing aeciospores that disperse to leaves of nearby shoots. Aeciospores on the upper leaf surfaces have been deposited from aecia on overhanging leaves.
cross fertilize to form aeciospores. These spores infect leaves of nearby plants in late spring, producing local lesions, which give rise to uredinia that, in turn, may infect other leaves. In the late summer the uredinia transform into telia that give rise to teliospores. In the late summer and fall, thistle plants that emerged in the spring senesce and diseased leaves with local lesions, that now contain telia, abscise, fall, and are blown onto newly emerging rosettes. Under conditions of favorable temperature and adequate dew, teliospores germinate into basidiospores that infect the rosettes. The fungus then develops hyphae that grow into the stem, and ultimately into the roots of the rosettes where the fungus survives the winter. Systemically diseased shoots emerge from this rootstock the following spring. What is significant in this disease cycle is: (1) the source of inoculum for systemic disease is teliabearing leaves from aeciospore infections in late spring and (2) the infection court is the newly emerging rosettes in the fall. The objective of this study was to test the validity of this disease cycle hypothesis by attempting to establish systemic disease, in fields in Kozani, Greece; Moscow, Russia; Christchurch, New Zealand; and Maryland, USA, through inoculation of rosettes of CT in the fall with telia-bearing leaves collected in mid-summer. 2. Materials and methods 2.1. Basic procedure
Fig. 3. Underside of leaf of Canada thistle bearing Puccinia punctiformis telia from aeciospore infections in late spring and early summer.
Despite research on the fungus dating to the work of Anton de Bary in 1863, the disease cycle has not been understood well enough to routinely establish the systemic state of the disease in disease-free thistle patches. However, from our own observations and research (Berner et al., 2012) and research reports in the literature (Bailiss and Wilson, 1967; Buller, 1950; Cockayne, 1915; Connick and French, 1991; Ferdinandsen, 1923; Frantzen, 1994a,b; French et al., 1987; French and Lightfield, 1990; Menzies, 1953; Olive, 1913; Rostrup, 1874; Van den Ende et al., 1987), there is enough information to form an hypothesis that is testable in the field. This hypothesis is as follows: Systemically diseased shoots arise, in the spring, from roots infected with the overwintering fungus. The first signs of the fungus on these shoots are orange spermagonia that
Prior to inoculations in the fall, disease-free fields infested with CT were located in each site in each location in Greece, New Zealand, Russia, and the USA. Except in the cases of inoculations with aqueous suspensions of teliospores, telia-bearing leaves were collected in mid- to late-summer in each country from areas that were remote from the intended test site. Leaves from each collection trip in each location were combined and checked for proportion teliospores by scraping, with a scalpel, pustules from six samples of leaves, from the collection of leaves, and counting teliospores and urediniospores from each sample. The leaves were allowed to dry in paper bags which were stored at ambient temperature, unless otherwise indicated, until the leaves were ready to be used. Immediately prior to use, the leaves were ground in a Waring blender. In the designed field tests (Section 2.3.) one-gram amounts of these ground telia-bearing leaves were weighed individually onto separate pieces of paper that were folded and stapled
D. Berner et al. / Biological Control 67 (2013) 350–360
to form individual packets. Individual packets were made for each inoculation of each rosette. In the designed field tests, colored flags on 1-m-long wires (or color-coded pegs), 20 each of four different colors, were mixed together, arbitrarily selected, and placed individually in the ground next to 40 randomly selected CT rosettes spaced a minimum of 2 m apart. The flag colors were coded for inoculation treatments of 2, 4, 6, and 8 multiple inoculations of each rosette. At the time of inoculation, each inoculated rosette was tagged with a plastic pot label corresponding in color to the color-coded number of inoculations and color-coded flags so that the appropriate rosette would be inoculated the correct number of times on successive inoculations. Inoculations were done every 3–4 days, with ground telia-bearing leaves (pre-packaged one-gram amounts), until all marked rosettes were inoculated the appropriate number of times, i.e., 2, 4, 6, or 8 times. Unless otherwise indicated, 10 rosettes were inoculated with each inoculation treatment. The fields were then left undisturbed, and no further inoculations were performed. 2.2. Initial field tests 2.2.1. Naturally occurring rosettes On October 9, 2008, about 4 ml (1 g) of ground telia-bearing leaf pieces was placed in the crown of each of 40–50 rosettes of CT in a disease-free field outside the USDA, ARS, Foreign Disesase-Weed Science Research Unit (FDWSRU) in Ft. Detrick, Maryland, USA, 39° 250 N; 77° 250 W. Rosettes that were inoculated had emerged in late August through September, 2008 from the bases of senescing CT shoots that had emerged in the spring of 2008. 2.2.2. Rosettes from transplants In June through July, 2009, CT leaves bearing telia of P. punctiformis were collected, stored, and allowed to dry as previously described. In July 2009, 976 one-month-old seedlings of CT that had been grown from cuttings in greenhouses at FDWSRU were transplanted into a 30 m 30 m area between two greenhouses at FDWSRU. All transplants were of a single clone. The area was tilled prior to transplanting, and no CT was previously found in the area. Transplants were spaced 1 m 0.75 m apart. One-half of the area was mowed in August, 2009 and allowed to re-grow. On September 25, 2009, about 4 ml (1 g) of ground telia-bearing leaf pieces was placed in the crowns of about half of the rosettes that had emerged from the bases of the original transplants. The other half of the rosettes was not inoculated. No further inoculations were performed.
353
per ml. The suspension was taken to the field, on a farmer’s property near Keymar, Maryland, 39° 370 N, 77° 160 W, and all marked rosettes were inoculated as described with 0.5 ml of the suspension. Twenty rosettes, 80 in total, were inoculated with each of the four inoculation treatments. The period of inoculations was from Sept. 10 to Oct. 6, 2010. 2.3.2. Inoculations with ground telia-bearing leaves 2.3.2.1. Kozani, Greece, 2010. A disease-free thistle-infested field was located near Kozani, Greece, 40° 180 N; 21° 470 E, in July 2010. Randomly selected CT rosettes were inoculated according to the basic procedure beginning on November 10 and continuing until December 1, 2010. 2.3.2.2. Keymar, Maryland, 2011. Telia-bearing leaves in this test were stored in paper bags at 80 C until ready for use. In August, 2011 a disease-free field infested with CT was found on a farmer’s property near Keymar, Maryland. This field was on the same farmer’s property but remote (about 500 m) from the field described in Section 2.2.2. The field was sub-divided into two sections, separated by about 30 m, one on a hillside and the other in a depression. In each section, inoculations of individual rosettes began on September 19, 2011 and continued until October 14, 2011. 2.3.2.3. Moscow, Russia, 2011. In August, 2011 a disease-free field infested with CT was found on the grounds of the All Russia Phytopathology Research Institute near Bolshie Vyazemy, Russia 55° 370 N; 36° 590 E, about 40 km from Moscow. The field was sub-divided into three areas, each separated by about 20 m, one in an open section, another in a shaded section, and the other in an area between these two sections. Beginning on September 21, and continuing until October 18, 2011, the rosettes in the open section were inoculated with ground leaves from Bolshie Vyazemy, and the rosettes in the other two sections were inoculated with ground telia-bearing leaves collected from Krasnodar, Russia 45° 10 N; 38° 580 E in June through August 2011. 2.3.2.4. Christchurch, New Zealand, 2012. In November 2011 (spring), four disease-free CT populations were selected within 10 km of Lincoln NZ in anticipation of autumn rust inoculations. Starting on 28 February (autumn) and continuing until March 20, 2012, 7 plants in each CT population were inoculated with each of the inoculation treatments. A total of 28 plants in each CT population were inoculated. 2.4. Data collection and statistical analysis
2.3. Designed field tests 2.3.1. Inoculations with aqueous suspensions of teliospores In 2010, systemically diseased shoots were produced in greenhouses at FDWSRU by inoculating adventitious shoot buds on root pieces of CT with an aqueous suspension of about 104 teliospores per shoot, according to the procedure of French and Lightfield, 1990, and then planting the inoculated root pieces into pots. About 6 weeks after planting, systemically diseased shoots began to emerge from the bases of the inoculated shoots. Teliospores from the inoculated plants were harvested, over several months, into gel caps with a micro-cyclone vacuum spore harvester (Tervet et al., 1951). The teliospores in gel caps were transferred to cryovials and stored at 80 C until the teliospores were required for use. On September 10, 2010, teliospores were removed from storage at 80 C and heat-shocked by submerging the cryo-vials containing the spores for 5 min in a 40 C water bath. Afterward, an aqueous suspension of teliospores was made by vortexing 40 mg teliospores in 480 ml of ultra-pure water plus 1 drop polysorbate 20, which produced a spore concentration of 2 106 teliospores
Data were collected each spring following the fall inoculations at each site. Data were collected when no further emergence of systemically diseased shoots occurred, about 60 days after the vernal equinox. In the initial field tests, data were simply the number and location of systemically diseased shoots emerged from the inoculated areas. In the structured field tests, the presence of one or more systemically diseased shoots emerging within 1 m from a color-coded flag/peg marking an inoculation treatment, i.e., 2, 4, 6, or 8 inoculations, was recorded as a positive reaction, a ‘‘1’’, for that flag/peg, i.e., previously inoculated rosette, in that inoculation treatment. The number of positive reactions for each inoculation treatment was totaled and divided by the number of inoculated rosettes to form proportions of positive reactions for each treatment in each site. Proportions of ‘‘0’’ and ‘‘1’’ were set equal to 0.01 and 0.99, respectively, and proportions were converted to logit values: logit = ln(proportion positive/(1 proportion positive)) for analyses. Statistical analyses were done only on data from the designed field tests. Each field, section, and year within a country was con-
354
D. Berner et al. / Biological Control 67 (2013) 350–360
sidered a repetition in the overall study. Data were analyzed, with the Mixed Procedure of SAS (SAS Institute Inc., 2004), as mixed models with repetitions as random effects and numbers of inoculations as fixed effects. Data were analyzed twice, once with inoculation times as a categorical variable and once with inoculation times as a continuous regression variable. For presentation, raw means of logit values were generated with the Means Procedure of SAS for each country and inoculation time. Estimated logit values from the analyses were back-transformed into percentages for ease in interpretation. 2.5. Weather data Year-long weather data were downloaded from the following named meteorological stations: Moscow (Russia), Kozani Airport (Greece), Frederick Municipal Airport (Frederick, Maryland, USA), Lincoln, Broadfield Ews Station (Christchurch, New Zealand). These stations were near the inoculation sites, and the weather data were downloaded for the respective years in which inoculations took place. Weather data included average, maximum, and minimum air temperatures, dew point temperatures, and precipitation. For some weather stations these data were daily averages while for other stations data were hourly observations. In the latter cases, data were read into SAS and daily averages, maxima, minima, and totals were calculated with the Means Procedure. Occurrence of dew on each day was estimated, in SAS, by subtracting minimum daily air temperature from average dew point temperature and regarding differences greater than zero as indicators (likelihood of dew) that dew occurred during that particular day. The ranges, means, and standard deviations of average temperatures and precipitation during the inoculation periods were calculated,
and likelihoods of dew were graphed versus day of year with SigmaPlot 10.0 (Systat software, Inc., Point Richmond, CA). 2.6. CT leaf abscission and new rosette emergence To assess the seasonal abscission of CT leaves, the number of leaves on each of 20 randomly selected CT shoots was counted weekly from July 5, 2012 until October 4, 2012 in the site described in Section 2.1.2. In the same site, the number of newly-emerged rosettes was sampled weekly during the same time period. Sampling for rosettes was done by throwing a hula-hoop, with an internal area of 0.385 m2, into different areas within the CT site 10 times, counting the number of rosettes in the hoop in each sample area, converting the rosette counts to number per m2, and calculating the average number of newly-emerged rosettes and the standard error among samples at each sampling date. The total number of leaves on the shoots and the average number of newly-emerged rosettes were graphed against sampling date, in days of year, with SigmaPlot 10.0.
3. Results 3.1. Initial field tests 3.1.1. Naturally occurring rosettes In the spring of 2009, a total of 12 systemically diseased CT shoots emerged from the area inoculated with telia-bearing leaves the previous October. No systemically diseased shoots were found outside the inoculated area. In the spring of 2010, no CT shoots, either healthy or diseased, emerged in this area.
Fig. 5. Site, at Ft. Detrick, Maryland, of Canada thistle inoculations, on September 25, 2009, with ground leaves bearing telia of Puccinia punctiformis. (A) Same site on May 5, 2011 with systemically diseased shoots marked with flags (B) Diagrammatic representation of site showing: prevailing wind direction, pre-inoculation mowing treatments, areas inoculated, systemically diseased shoots (3, orange text and lines) emerged in May 2010, systemically diseased shoots (160 shoots in 70 clusters, blue text and lines) emerged on May 5, 2011, and systemically diseased shoots (22 shoots in 12 clusters, purple text and lines) emerged on May 3, 2012. (C) Same site on May 3, 2012 showing areas devoid of Canada thistle, stunted shoots, and newly-emerged systemically diseased shoots (22 shoots in 12 clusters, blue flags photo C). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
355
D. Berner et al. / Biological Control 67 (2013) 350–360
3.1.2. Rosettes from transplants In the spring of 2010, three systemically diseased CT shoots emerged from the area inoculated with telia-bearing leaves the previous September. In the spring of 2011, 160 systemically diseased CT shoots in 70 clusters emerged in this area (Fig. 5A and B). Most of the emerged diseased shoots were in the section that had not been previously mowed and were downwind from the prevailing wind direction. In the spring of 2012, an additional 22 systemically diseased shoots in 12 clusters emerged. By spring 2012, thistle density had declined, and many of the remaining shoots in the area with the most systemically diseased shoots in spring 2011 were severely stunted (Fig. 5C).
different than zero in t-tests against zero, but there were no significant differences among the inoculations. Results of mixed model analysis with inoculation times as a continuous regression variable are presented in Fig. 6 that includes the ANOVA table and means of logit values for the different inoculations in different countries. The slope for number of inoculations was not significant, but it was positive, indicating a slight increase in proportions of systemic disease with increased number of inoculations. There was no definite pattern associated with number of inoculations within countries based on means of logit values.
3.2. Designed field tests
The ranges, means, and standard deviations of average temperatures during the inoculation periods were: Ft. Detrick, Maryland, Sept. 10-Oct. 6, 2010, 12.1–27.0 °C, mean 19.3 ± 3.40; Ft. Detrick, Maryland, Sept. 19-Oct. 14, 2011, 7.3–22.6 °C, mean 16.5 ± 4.33; Kozani, Greece, Nov. 10-Dec. 1, 2010, 6.1–16.1 °C, mean 11.3 ± 2.42; Moscow, Russia, Sept. 21-Oct. 18, 2012, 2.3–16.7 °C, mean 8.9 ± 3.50; Christchurch, New Zealand, Feb. 28-March 20, 2012, 9.1–19.3 °C, mean 13.4 ± 2.78. Temperatures gradually declined in all countries following the inoculation periods. The temperature range for P. punctiformis teliospore germination is reported to be between 8 and 25 °C, with maximum germination between 16 and 20 °C (French and Lightfield, 1990). Frantzen (1994c) concludes that the optimum rate of teliospore germination, which may result in more infections per unit time, occurs at 10–15 °C. Daily precipitation ranges, means, and standard deviations during these inoculation periods were: Ft. Detrick, Maryland, 2010 (range 0–139 mm, mean 8.6 ± 27.35 mm); Ft. Detrick, Maryland, 2011 (range 0–115.6 mm, mean 8.7 ± 22.69 mm); Kozani, Greece (range 0–11 mm, mean 1.3 ± 3.26 mm); Moscow, Russia (range 0–26 mm, mean 2.1 ± 5.17 mm); Christchurch, New Zealand (range 0–31 mm, mean 2.6 ± 6.90 mm). There was a considerable range in precipitation among locations: 1.3–8.7 mm mean daily precipitation and 11–139 mm total daily precipitation during the inoculation periods.
The proportion of teliospores used for aqueous inoculations at Keymar, Maryland in 2010 was approximately 50%. Proportions of teliospores in samples of telia-bearing leaves for the other inoculations were: Keymar, Maryland 2011 P70%; Kozani, Greece 2010 P80%; Bolshie Vyazemy, Russia, 2011 spores from Bolshie Vyazemy P65%; Bolshie Vyazemy, Russia, 2011 spores from Krasnodar P50%; all New Zealand sites 2012 P20% (resulted in about 3.8 106 teliospores ml1 when examined with a haemocytometer). Percentages of inoculated rosettes that gave rise to at least one systemically diseased shoot in the spring following inoculations are presented in Table 1. There was considerable variability among inoculation treatments within and among field sites, but at least 10% of rosettes in all inoculated sites gave rise to at least one systemically diseased shoot. The highest percentages occurred in New Zealand, but these percentages were based on seven inoculated rosettes, as opposed to 10 or 20 inoculated rosettes in other countries. Mixed model analysis of logit values, with repetitions as random effects and inoculation times as categorical fixed effects, resulted in the following logit estimates for the different number of inoculations (back-transformed percentages are in parentheses): 2x = 1.9832 (12.1%); 4x = 1.2473 (22.3%); 6x = 1.2942 (21.5%); 8x = 1.3212 (21.1%). All estimates were significantly (P 6 0.05)
3.3. Weather
Table 1 Percentages of inoculated rosettes giving rise to at least one systemically diseased shoot in the spring following inoculations of rosettes the previous fall. Individual rosettes were inoculated 2, 4, 6, or 8 times with either a suspension of teliospores or about 1 g of telia-bearing leaves. Site/inoculation year
Percent of inoculated rosettes giving rise to a systemically diseased shoot Keymar, Maryland, USA, 2010a Keymar, Maryland, USA, 2011 Hilltop siteb Keymar, Maryland, USA, 2011 Depression siteb Kozani, Greece, 2010b Bolshie Vyazemy, Russia, 2011 Open field siteb – telia bearing leaves from Bolshie Vyazemy, Russia Bolshie Vyazemy, Russia, 2011 Shaded field siteb – telia bearing leaves from Krasnodar, Russia Bolshie Vyazemy, Russia, 2011 Intermediate field siteb – telia bearing leaves from Krasnodar, Russia Christchurch, New Zealand, 2012 Moir sitec Christchurch, New Zealand, 2012 Boyle Toitoi sitec Christchurch, New Zealand, 2012 Boyle Belam sitec Christchurch, New Zealand, 2012 Lu sitec Mean a b c
Number of inoculations 2
4
6
8
Mean
0 20
0 30
10 40
0 50
2.5 35.0
0
10
10
10
7.5
0 30
40 40
10 50
40 10
22.5 32.5
0
10
0
0
2.5
50
10
20
30
27.5
85.7
71.4
28.6
42.9
57.2
14.3
42.9
57.1
85.7
50.0
42.9
57.1
57.1
42.9
50.0
28.6
14.3
28.6
28.6
25.0
24.68
29.61
28.31
30.92
Twenty rosettes inoculated per each inoculation treatment with aqueous suspensions of 106 teliospores. Ten rosettes inoculated per each inoculation treatment with about 1 g of ground telia-bearing leaves. Seven rosettes inoculated per each inoculation treatment with about 1 g of ground telia-bearing leaves.
356
D. Berner et al. / Biological Control 67 (2013) 350–360
Fig. 6. Results of mixed model logistic regression of number of inoculations with ground leaves bearing telia of Puccinia punctiformis versus logit values of proportion systemically diseased shoots emerged adjacent to sites of inoculated rosettes. Countries, field sites, and years were repetitions in the analysis and were treated as random effects with number of inoculations as a fixed effect. The regression line is based on parameters from the mixed model ANOVA. Means of logit values for each number of inoculations in each country are also plotted.
Fig. 7. Daily likelihoods of dew based on average dew point and minimum air temperature records collected from named meteorological stations near inoculation sites in the respective years of inoculations. Likelihood of dew was calculated daily as average dew point temperature minus minimum air temperature.
D. Berner et al. / Biological Control 67 (2013) 350–360
Estimated likelihoods of dew for the years of inoculations in each country are presented in Fig. 7. In each country there was an extended period, in the fall, when dew was most likely to occur, as indicated by the prolonged periods above zero, calculated by average dew point temperature minus minimum air temperature. These periods are evident on the graphs and are delimited by vertical lines. The calendar dates corresponding to these days of the year are: August 30–September 29 in Moscow, Russia; September 29–December 1 in Kozani, Greece; September 5–October 13 in Frederick, Maryland; and March 16–April 28 in Christchurch, New Zealand. Inoculations began either within these dates or immediately before them. Free moisture (dew) and favorable temperatures are two primary requirements for teliospore germination; both of these conditions were present in the fall during inoculations in all countries.
357
3.4. CT leaf abscission and new rosette emergence Results of the sampling studies on timing of abscission of CT leaves and emergence of new rosettes is shown in Fig. 8. Leaf abscission was gradual until day 240 (August 27), when leaves begin to fall at an accelerated rate, until day 278 (October 4), when all leaves had abscised. Emergence of new rosettes accelerated between day 249 and 256 (September 5–12) and accelerated emergence coincided with rapid leaf abscission. A simple correlation between number of emerged rosettes and the total leaves remaining on the shoots produced a Pearson correlation coefficient of 0.867 with a P > |r| of 0.0012. These periods of rapid leaf abscission and accelerated rosette emergence also coincided with dates of prolonged dew in 2011 (September 5–October 13) and with dates of prolonged dew (August 27–October 15) recorded from 2003–2008 at Frederick airport. Senescing CT shoots and newlyemerging rosettes underneath these shoots are shown in Fig. 9. The Fig. 9 photo is from Keymar, Maryland on September 16 2011, about the time of inoculations in that site and within the period of prolonged dew. 4. Discussion
Fig. 8. Total number of leaves, counted weekly, on shoots of 20 randomly selected Canada thistle shoots and weekly averages and standard errors of number of newlyemerged rosettes from 10 random samples. Leaf and rosette counts were made in the same field at Ft. Detrick, Maryland in 2012.
The validity of the P. punctiformis disease cycle proposed in the introduction has been demonstrated by establishing systemic disease, in all 13 CT infested fields in four countries, through inoculation of CT rosettes in the fall with either teliospores or telia-bearing leaves collected in mid-summer. Supporting evidence for the validity of this disease cycle comes from weather data collected from the four countries that show temperature and dew conditions are most favorable for teliospore germination during the respective fall seasons. Additional evidence that telia-bearing leaves are the natural inoculum and that newly-emerging rosettes are the natural receptive infection courts comes from the sampling study that shows coincidence of CT leaf abscission in the late summer and fall with emergence of new rosettes at a period of favorable temperature and dew for teliospore germination. CT leaves bearing telia from P. punctiformis aeciospore infections in late spring and early summer naturally abscise or are blown onto newly emerging rosettes during this favorable environmental period in the fall. (This
Fig. 9. Senescing Canada thistle shoots and leaves overhanging newly emerging rosettes in Keymar, Maryland, September 16, 2011.
358
D. Berner et al. / Biological Control 67 (2013) 350–360
appeared to be the case at the Ft. Detrick site (Fig. 5A and B) where movement from initially established systemically diseased shoots in 2010 was predominately downwind and resulted in a tremendous increase in number of systemically diseased shoots in 2011). Teliospores within telia-bearing leaves then germinate into basidiospores which, under the same favorable environmental conditions, infect the rosettes and establish the fungus to overwinter in root systems. The CT leaf abscission and new rosette emergence study should be repeated in other areas and countries, but demonstration of establishment of systemic disease with telia-bearing leaves inoculated onto rosettes in the fall strongly supports this scenario. Recently, transmission of urediniospores of P. puunctiformis by a stem-mining weevil (Ceratapion onopordi) has been proposed as the mechanism and spore type necessary for systemic infection of CT in the field (Wandeler and Bacher, 2006; Wandeler et al., 2008). However, Cripps et al. (2009) showed, on a country-tocountry basis, that systemic rust of CT occurs independent of the presence of the weevil. Muller et al. (2011) injected CT stems with P. punctiformis urediniospores (plus
