Maliau bracket fungi project

From BioDivBorneo2010

Jump to: navigation, search

Reproductive ecology and circadian rhythms of Ganoderma sp. bracket fungi in Maliau Basin Conservation Area, Malaysia

Rachel Hawkins and Cameron Kirk-Giannini

Contents

Abstract

Basidiomycete fungi, especially saprophytic bracket fungi, are ecologically and environmentally important but understudied plant pathogens. In this study, we investigate the reproductive ecology of a tropical species in the basidiomycete genus Ganoderma, paying special attention to rates of spore dispersal in reproductively active fungi. A solid-substrate spore trapping technique was utilized, along with frequent microclimatological measurements, to determine the rates of spore dispersal in three individual fungi located within the Maliau Basin Conservation Area in Sabah, Malasia, and test whether rates of spore dispersal varied with one or more of temperature, relative humidity, light intensity, and wind speed. Spores were collected and measurements taken at two-hour intervals during daylight and early evening hours over a period of two days, and statistical analysis was conducted using the R programming environment. Our results do not indicate that any of the tested factors affect the rate of spore dispersal in our three fungi. Our estimates of absolute rates of spore dispersal per fertile surface area, moreover, are low compared to similar studies in the literature, possibly indicating a low spore capture rate for solid-substrate trapping techniques. We conclude that further study of a larger population of dispersing fungi will be required to satisfactorily determine which factors, if any, influence rates of diurnal spore dispersal, and that an air trap will be necessary to accurately measure total numbers of spores dispersed.

Introduction

Ganoderma is a cosmopolitan genus of saprophytic basidiomycete bracket fungi which feed on decaying wood and debris. Various Ganoderma species have attracted scientific attention in recent years because of their important ecological role as carbon recyclers and
Figure 1: Fungi individuals 1-3 on decaying tree substrate.
because several are important pests of economically relevant cultivated tree species, especially the oil palm (genus Elaeis).[1] Most studies have been directed toward the reproductive ecology of Ganoderma species, especially at determining the quantity of spores release during reproduction and elucidating diurnal patterns of spore dispersal. As a result, a growing literature exists regarding the reproductive ecology of Ganoderma bracket fungi in temperate climates.[2][3] Nevertheless, despite the fact that the ecological and agricultural impacts of Ganoderma are most pronounced in tropical regions, no comparable studies have yet been performed in tropical latitudes or areas of extensive oil palm production. While patterns of Ganoderma spore dispersal have been found to correspond with changes in humidity and temperature in temperate zones[4], tropical saprophytes may have different adaptations to the drastically different climates.

We investigate the diurnal spore dispersal of three individuals of a tropical Ganoderma species found in the Maliau Basin Conservation Area of Sabah, Malaysia (see Figure 1). The fungi were observed visibly dispersing spores late in the morning (11:00 a.m.) of 28/6/2010 with no visible associated mechanical disturbance. Since active dispersal during daylight is uncommon,[5] measurements of spore dispersal as well as local temperature, relative humidity, wind speed, and light intensity were begun immediately and continued over a 48-hour period. The discovery of such an uncommonly active population of Ganoderma individuals presents an important opportunity to better understand the reproductive ecology of the tropical representatives of this genus.

Questions

We seek to answer three questions:

  • Which factors (if any) among temperature, relative humidity, wind speed, and light intensity affect the temporal distribution of the rate of the fungi's spore dispersal?
  • To what extent do these factors affect said temporal distribution?
  • What can be said about the total daily dispersal of the spores based on measurements of diurnal dispersal?

Methods

Data Collection

Spores were collected and environmental parameters (temperature, relative humidity, wind speed, light intensity) measured every two daylight hours for two consecutive days (0600 to 1800 on 6-30-2010; 0800 to 2000 on 7-1-2010). Observation hours were staggered over the two-day period to encompass a greater portion of total diurnal variation: 0600 to 2000. During each measurement, a spore trap was held at a distance of 2cm from the fertile surface of each fungus for a duration of 3 minutes. Simultaneously, the humidity, temperature, wind speed and light intensity of each fungal microhabitat were measured using a portable meteorological multimeter (Kestrel). The times and durations of rainstorms during daylight hours were also recorded.

Figure 2: Spore slides prepared for counts and analysis. Cover slip grid visible near center.

Spore traps were composed of a washed petri dish used to support and protect an adhesive substrate (clear plastic packing tape). Traps were assembled in a remote location and kept enclosed until the time of measurement. Trapped spores were transferred to glass microscope slides, leaving the adhesive substrate intact.

Spore Counts

Spores were observed at 100x magnification using a light microscope. An arbitrary 2.66 mm<math>^2</math> area was defined by placing a standard four-quadrant gridded coverslip in the center of each slide (see Figure 2). Two observers independently recorded the number of spores in the observation area, and the numbers were then combined to compensate for observer bias. To determine the uniformity of the spore distribution across each slide, three slides were chosen using a randomization procedure and each was measured in three separate locations (left, middle, and right) by a single observer. Small variance values in these measurements demonstrated the uniformity of the distribution of the spores on the microscope slides.

Figure 3: Grid-paper tracings of fungi.

Fungal Impressions

The shapes of the fungi were also recorded. Graph paper was placed firmly against the fertile surface of each fungus, and the outlines of the edges were traced. For fungi too large to fit onto one sheet of paper, two sheets were overlapped, and their points of overlap were marked and later combined into one outline (see Figure 3).

Analysis

Factors Affecting Spore Dispersal

All data were analyzed in the R statistical programming environment. Raw data imported from resource files was stored in a dataframe, and raw spore counts from the two observers were added according to the spore count method above. Plots of spore counts for each fungus on each day with respect to each independent variable were generated, organized into sets, and written to external files in the sink directory.

Because the spore counts from the three fungi were of different orders of magnitude, and because the motivating hypotheses of our study relate to cyclical variations rather than absolute values, the spore count data for each fungus over the 2-day period was normalized according to the formula:

<math>\frac{x - mean(x)}{sd(x)}</math>

General linear models were constructed using a poisson family distribution in order to analyze the relationship between spore counts and the independent variables of temperature, relative humidity, light presence, and wind speed. For each set of variables we performed the general linear model twice: one test considered the independent variable without regard to the individual fungi, the other considered the independent variable with regard to the individual fungi. The following general linear models were constructed from the non-normalized data:

  1. Spore counts in terms of relative humidity
  2. Spore counts in terms of temperature
  3. Spore counts in terms of light presence
  4. Spore counts in terms of wind speed
  5. Spore counts in terms of relative humidity and individual fungi
  6. Spore counts in terms of temperature and individual fungi
  7. Spore counts in terms of light presence and individual fungi
  8. Spore counts in terms of wind speed and individual fungi


The following scatterplots were generated to visualize relationships in the data sets that appeared most promising:

  1. A scatterplot of total spore counts against temperature
  2. A scatterplot of total spore counts against humidity
  3. A scatterplot of normalized spore counts against temperature
  4. A scatterplot of normalized spore counts against humidity


These scatterplot results suggested a possible relationship between normalized spore counts and temperature. To determine whether there was any cyclical pattern in either dataset, two boxplots were generated comparing normalized spore counts and temperature, respectively, at each measured time point (see Figure 6).

Additionally, two t-tests were used to determine whether the spore counts and temperature at each measured time point were distributed non-randomly. For each data subset, the measured time points 1000, 1200, 1400, and 1600 were tested against the remaining measured time points (0600, 0800, 1800, and 2000).

Fungal Impressions

Figure 4: Fungal impression of Fungus 2 approximated as an ellipse.

To estimate the area of the underside of the fungus, the area traced onto the graph paper was approximated as an ellipse (see Figure 4) using the following formula:

<math>A=\pi\,\!*a*b</math>

(where <math>a=0.5*length of major axis</math> and <math>b=0.5*length of minor axis</math>).

Positions of axes were approximated visually and made as symmetrical as possible according to the outline of each fungus.

The two observer spore counts for each fungus at each time point were themselves averaged, yielding a single count value for each time point in the two-day observation period, the values thus obtained were averaged according to day, and these mean numbers from each day were finally averaged together to get one approximate two-day value to use for extrapolation.

Figure 5: "Figure 1" from Sreeramulu (1963)

Extrapolations were conducted using data collected by Sreeramulu (1963): Figure 5 shows his chart entitles "diurnal periodicity of Ganoderma applanatum spores expressed as percentage of the peak geometric mean concentration"[6] Figure 5 was used as a model for the fold-difference between rates of midday spore release and rates nocturnal peak spore release, and this model was used to calculate theoretical nocturnal peak spore release rates for the three fungi measured. Midday points of Sreeramulu's dtata were averaged; this mean number was then used to calculate the factor (7.7) by which spore numbers were multiplied to reach an approximate nocturnal peak value for each fungus. Average spore dispersal numbers per cm<math>^2</math> were then calculated for midday values and for extrapolated nocturnal peaks.

Results

Factors Affecting Spore Dispersal

Figure 6: Time-dependent patterns in temperature and adjusted count patterns

All general linear model tests reported in the Analysis section above, with the exception of those including wind-speed data, yielded highly significant results (p-values on the order of 2 x 10<math>^-</math><math>^1</math><math>^6</math>). However, in most cases, the slope of the associated fitted linear curve approximated m=0, raising the suspicion that the significance of these p-values indicated the clustering of our count data around a single value which was insensitive to the tested independent variable. This hypothesis was pursued by visualizing the data with scatterplots, which made it clear that the divergence of our slope values from m=0 was due to the presence of outliers distorting the fitted curves.

The box plot of the normalized count data with respect to time of day, on the other hand, appeared to exhibit a cyclical diurnal pattern in count data with a midday peak. This pattern mirrored the strong diurnal pattern revealed in a similar box plot of temperature data with respect to time of day (see Figure 6). To test the robustness of the observed cyclical diurnal patterns, t-tests were conducted comparing midday values (datapoints at t = 1000, 1200, 1400, and 1600) with morning and evening values (datapoints at t = 0600, 0800, 1800, and 2000). These t-tests returned p-values of 4.24 x 10<math>^-</math><math>^8</math> and 0.5822 for temperature and normalized count data, respectively.

Fungal Impressions

Table 1: Calculated spore counts from fungal impressions (per cm<math>^2</math>)

The 3 spore-releasing fertile surfaces of the fungi were of comparable sizes: Fungus 1 had an area of 869 cm<math>^2</math>; Fungus 2, 900 cm<math>^2</math>; and Fungus 3, 970cm<math>^2</math>. Total spore counts per cm<math>^2</math> tended to vary by an order of magnitude between measured diurnal values and extrapolated nocturnal values. The values calculated for Fungus 2 were an order of magnitude larger than their respective counterparts for Fungus 1 and Fungus 3. However, the higher values for Fungus 2's spore release do not seem to be results of the size of the fertile surface of the fungus.

Table 1 lists the average spore totals calculated using the recorded diurnal spore counts and extrapolated noctural spore predictions based on Sreeramulu (1963).

Discussion

Factors Affecting Spore Dispersal

As stated previously in the Results section, we conclude that the p-values attained from the general linear models, although highly significant, do not display a strong relationship between the independent variables (relative humidity, temperature, light presence, and wind speed) and the non-normalized spore counts. The conjunction of very low slope values and extremely low p-values seems to be caused by the presence of outliers and count data which do not fluctuate with regard to the independent variables.

The t-tests which were performed on both temperature and normalized spore count data, each with respect to measured time points, yielded disparate p-values. The t-test of temperature vs. time yielded a p-value of 4.238 x 10<math>^-</math><math>^5</math>, which is highly significant. The t-test of normalized spore count data yielded a p-value of 0.5822, which is not significant. The stark difference between the p-values shows that there is no evidence that there is a relationship between midday normalized count data morning/evening normalized count data: if there is no robust cyclical pattern in normalized spore counts, as indicated by the high p-value of our t-test, then a fortiori there is no cyclical correspondence of normalized spore counts with temperature.

Fungal Impressions

It appears that the size of the fungus does not correspond to total spore counts within our sample, as evidenced by Fungus 2, which has the highest spore counts but not the largest fertile area. Our calculated spore totals are relatively low compared to the literature, which makes estimates of spore dispersal comparable to 3.1 x 10⁷ spores/hr/cm².[7]

Possible Sources of Error

Many sampling methods for analysis of basidiomycete spore counts involve sophisticated, automated air samplers which collect data continuously. Rockett and Kramer (1974) suggest that adhesive-based solid-substrate spore collection traps are not sufficient to handle the vast number of spores released by these fungi.[8] The comparatively lower numbers of spore counts calculated from the fungal impressions may be an artifact our our sampling method rather than differences in the amount of spore dispersal.

The time-limited nature of the experiment may also have not yielded as strong of a data set as possible. Having the ability to sample at least every hour over a 24-hour period for a longer span of time would give a more robust data set which would be more likely to show any significant relationships of spore counts to independent variables.

Estimation processes for the fungal impressions may have distorted the spore counts; however, all of these data would differ by the same coefficient if such were the case. Because the data is being queried primarily for relationships among its own points, this is not a methodological problem.

Directions for Further Study

As important and understudied members of the tropical ecosystem in the Maliau Basin, Sabah, Malaysia, tropical saprophytes, particularly Ganoderma, ought to be studied in order to understand their effects on carbon cycling and economic development as well as the character of their reproductive ecology. Nocturnal sampling in addition to diurnal sampling of spores ought to be carried out to fully understand the circadian cycles of spore dispersal. Because Ganoderma is also a pest of the oil palm, sampling of spores in oil palm production areas as well as in primary and secondary forests ought to be carried out. Knowing how this "pest" fungus spreads within or between each ecosystem will be valuable information to consider in landscape planning as Sabah becomes increasingly economically developed and seeks to protect its wild areas.

References

  1. Paterson, R.R.M. "Ganoderma disease of oil palm—A white rot perspective necessary for integrated control."Crop Protection 26 (2007) 1369–1376.
  2. Sreeramulu, T. "Observations on the Periodicity in the Air-Borne Spores of Ganoderma applanatum." Mycologia, Vol. 55, No. 4 (Jul. - Aug., 1963), pp. 371-379.
  3. Kramer, C.L. and Long, D.L. "An endogenous rhythm of spore discharge in Ganoderma applanatum." Mycologia, Vol. 62, No. 6 (Nov. - Dec., 1970), pp. 1138-1144.
  4. Ibid.
  5. Pringle, A. Personal communication. Jun 2010.
  6. Op. cit.
  7. Rockett, T.R and Kramer, C.L. “Periodicity and total spore production by lignicolous basidiomycetes.” Mycologia, Vol. 66, No. 5 (Sep. - Oct., 1974), pp. 817-829.
  8. Ibid.

Appendix

R Scripts

# Prepare Workspace
rm(list=ls())
setwd("Documents/Current/School/BB10/mrfiles")
sink("Output/outfile")

# Import Data
fd1 <- read.table("fd1.csv", sep = "|", header = T)
fd2 <- read.table("fd2.csv", sep = "|", header = T)
h1 <- read.table("h1.csv", sep = "|", header = T)
h2 <- read.table("h2.csv", sep = "|", header = T)
t1 <- read.table("t1.csv", sep = "|", header = T)
t2 <- read.table("t2.csv", sep = "|", header = T)
l1 <- read.table("l1.csv", sep = "|", header = T)
l2 <- read.table("l2.csv", sep = "|", header = T)
w1 <- read.table("w.csv", sep = "|", header = T)
w2 <- read.table("w.csv", sep = "|", header = T)

# Add Spore Counts
fd1$f1 <- ((fd1$f1r + fd1$f1n))
fd1$f2 <- ((fd1$f2r + fd1$f2n))
fd1$f3 <- ((fd1$f3r + fd1$f3n))
fd1 <- fd1[,-(1:6)]
fd2$f1 <- ((fd2$f1r + fd2$f1n))
fd2$f2 <- ((fd2$f2r + fd2$f2n))
fd2$f3 <- ((fd2$f3r + fd2$f3n))
fd2 <- fd2[,-(1:6)]

# Make Master Chart
fun<-c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3)
hum<-c(h1$f1,h2$f1,h1$f2,h2$f2,h1$f3,h2$f3)
tmp<-c(t1$f1,t2$f1,t1$f2,t2$f2,t1$f3,t2$f3)
lt<-c(l1$f1,l2$f1,l1$f2,l2$f2,l1$f3,l2$f3)
wnd<-c(w1$f1,w2$f1,w1$f2,w2$f2,w1$f3,w2$f3)
ct<-c(fd1$f1,fd2$f1,fd1$f2,fd2$f2,fd1$f3,fd2$f3)
mstr<-data.frame(fun,hum,tmp,lt,wnd,ct)
mstr

# Adjust Master Chart
adjct<-c((c(fd1$f1,fd2$f1)-mean(c(fd1$f1,fd2$f1)))/sd(c(fd1$f1,fd2$f1)),(c(fd1$f2,fd2$f2)-mean(c(fd1$f2,fd2$f2)))/sd(c(fd1$f2,fd2$f2)),
(c(fd1$f3,fd2$f3)-mean(c(fd1$f3,fd2$f3)))/sd(c(fd1$f3,fd2$f3)))
data.frame(mstr$fun,
mstr$hum,
mstr$tmp,
mstr$wnd,
mstr$ct,
adjct) -> adjmstr
adjmstr$time <- c(600,800,1000,1200,1400,1600,1800,800,1000,1200,1400,1600,1800,2000,600,800,1000,1200,1400,1600,1800,800,1000,1200,1400,1600,1800,2000,
600,800,1000,1200,1400,1600,1800,800,1000,1200,1400,1600,1800,2000)
colnames(adjmstr) <- c("fun","hum","tmp","lt","wnd","ct","time")
adjmstr

# Plot Imported Data
jpeg(file = "Output/plot%d.jpeg", width = 500, height = 750)

par(mfrow=c(3, 2))
plot(fd1$f1, type = "l")
plot(fd2$f1, type = "l")
plot(fd1$f2, type = "l")
plot(fd2$f2, type = "l")
plot(fd1$f3, type = "l")
plot(fd2$f3, type = "l")

par(mfrow=c(3, 2))
plot(h1$f1, type = "l")
plot(h2$f1, type = "l")
plot(h1$f2, type = "l")
plot(h2$f2, type = "l")
plot(h1$f3, type = "l")
plot(h2$f3, type = "l")

par(mfrow=c(3, 2))
plot(t1$f1, type = "l")
plot(t2$f1, type = "l")
plot(t1$f2, type = "l")
plot(t2$f2, type = "l")
plot(t1$f3, type = "l")
plot(t2$f3, type = "l")

par(mfrow=c(3, 2))
plot(l1$f1, type = "l")
plot(l2$f1, type = "l")
plot(l1$f2, type = "l")
plot(l2$f2, type = "l")
plot(l1$f3, type = "l")
plot(l2$f3, type = "l")

par(mfrow=c(3, 2))
plot(w1$f1, type = "l")
plot(w2$f1, type = "l")
plot(w1$f2, type = "l")
plot(w2$f2, type = "l")
plot(w1$f3, type = "l")
plot(w2$f3, type = "l")

# Scatterplots of Count and Adjusted Count vs. Humidity and Temperature

par(mfrow=c(2, 2))
plot(mstr$ct ~ mstr$hum)
plot(mstr$ct ~ mstr$tmp)
plot(adjmstr$ct ~ adjmstr$hum)
plot(adjmstr$ct ~ adjmstr$tmp)

# Boxplots of Adjusted Count and Temperature vs. Time

par(mfrow=c(2, 1))
boxplot(tmp~time, data = adjmstr, main = "Time-Dependent Patterns in Temperature and Adjusted Count Values", xlab = "Time (h)", ylab = "Temperature (°C)")
boxplot(ct~time, data = adjmstr, xlab = "Time (h)", ylab = "Adjusted Counts (Standard Deviations from the Mean)")

# Perform glm Tests
summary(glm(mstr$ct ~ mstr$hum, family = poisson))
summary(glm(mstr$ct ~ mstr$tmp, family = poisson))
summary(glm(mstr$ct ~ mstr$lt, family = poisson))
summary(glm(mstr$ct ~ mstr$wnd, family = poisson))

summary(glm(mstr$ct ~ mstr$hum + mstr$fun, family = poisson))
summary(glm(mstr$ct ~ mstr$tmp + mstr$fun, family = poisson))
summary(glm(mstr$ct ~ mstr$lt + mstr$fun, family = poisson))
summary(glm(mstr$ct ~ mstr$wnd + mstr$fun, family = poisson))

Files for Download

  1. Complete outfile of our R script. (PDF)
  2. Complete collection of figures produced by our R script. (PDF)