NATURAL ENEMY EVALUATION · biological control in crop IPM ... Slides of David Biddinger, ... Cage exclusion example #2: control of cassava mealybug

NATURAL ENEMY

EVALUATION

Read ch 20

INCORPORATION

NATURAL ENEMIES INTO

IPM DECSION MAKING

Requirements for inclusion of

biological control in crop IPM

Effective BC agents must exist

Target must be important relative to other crop pests

Lack of major conflict with control of key pests

Tolerance of some level of target pest in crop

Realistic economic injury level

Sound understanding of natural enemy’s ecology

Sampling tools to measure natural enemy abundance

Modification of IPM thresholds to reflect information

about current natural enemy densities or ratios in the crop

Monitoring tomato fruitworm (Helicoverpa zea) in processing

tomatoes in CA. Ratios of black (parasitized)/white (healthy) eggs

are used in a sequential sampling plan to assess need for pesticides

Modification of IPM thresholds to reflect information

about current natural enemy densities or ratios in the crop

Modification of IPM thresholds-second example

The pest threshold of apple blotch leafminer (13 mines/100

leaves of 1st gen.) can be relaxed in view of 1st gen larval

parasitism, which can be determined by timely sampling

Monitoring Southern Pine Beetle and its Key

Predator to Decide if “Spots” Must be Controlled

SOUTHERN PINE BEETLE PREDICTION CHART

(Modified January 2002)

Average Number

of SPB/Trap/Day

Severe Outbreak

Increasing or High

Static or Moderate

Declining or Low

Average Percent SPB = #SPB x 100/ (#SPB + # Clerids)

55%

20%

6%

## 10% 20% 30% 80% 90% 100%40% 50% 60% 70%

0

10

20

30

80

40

50

60

70

Control of mites in PA apple orchards

Integration of predator counts with mite counts

Two historical eras:

a) When Stethorus punctum was the key predator

b) When it was replaced by Typhlodromus pyri, a

predatory mite

Slides of David Biddinger, Pennsylvania State University, with modification

European Red Mite Two-Spotted Spider Mite

Stethorus punctum Reduced miticide use by

50% in PA from late 1970s to mid 90s.

Saved growers $20 million over 15 year period.

Reduced miticide use by over 2.2 million pounds.

Required at least 5 mites/leaf to stay in orchard and

8-10 mites/leaf to begin laying eggs.

Stethorus punctum

– Some leaf injury (bronzing)

often occurred.

– Important when:

• methomyl was necessary to

control TABM - predator mites

were not a factor.

• Few miticides available & not

very effective due to resistance.

• Tree were large, industry was

predominantly processing.

Adult & larva

DJB-2005

When Stethorus punctum made sense

– Impact of new insecticides that were very toxic:

• Neo-nicitinoids (imidacloprid)

• Chitin inhibitor IGRs (novaluron, diflubenzuron, buprofezin)

• Abamectin miticide

– More effective miticides registered.

– Mite thresholds/tolerance for injury declined.

• Move to smaller trees.

• More suceptible varieties.

• Shift from processing to fresh market where color was important.

ERM & Stethorus Eggs

DJB-2005

Mite Predators Changed

Typhlodromus pyri

Stethorus punctum

Disappeared due

to new pesticides

New predator invaded region and was both

able to withstand the new pesticides and

provided better mite control

Monitoring Predator Mite and Pest Mite

Numbers Guides Spray Decisions

Pest mite to T. pyri ratios of 10 to 1 or lower are sufficient for biological control

ASSESSING IMPACT OF

CLASSICAL BIOLOGICAL

CONTROL AGENTS

ASSESSING IMPACT OF CLASSICAL BIOLOGICAL

CONTROL AGENTS

Experimental methods based on “with

and without” treatment plots

1. Before and after plot

2. Geographic release and control plots

3. Cage exclusion

4. Chemical exclusion

ASSESSING IMPACT OF

CLASSICAL BIOLOGICAL CONTROL AGENTS

METHOD #1.

“BEFORE AND AFTER”

ASSESSMENT OF PEST DENSITY

OR CROP LOSS

Example #1

Invasion of

citrus blackfly

into Florida

and its control

via parasitoid

introduction

Black= nymphs

Citrus blackfly (nymphs are black normally), with exit

holes of introduced parasitoids

Citrus blackfly parasitoid: Amitus hesperidum, one of two

species introduced for pest control in Florida

Sampling to measure density and parasitism

before and after introduction

Yard citrus is main pest reservoir

Citrus blackfly nymphal density “before and

after” parasitioid introduction

Post-project pest density

Citrus blackfly nymphal density

“before and after” parasitioid introduction

Pre-project pest density

Transition phase

as parasitism

rises

CMF

A. hesperidus

P. opulenta

Example #2

Olive scale,

Parlatoria

oleae in CA,

before

parasitoid

introduction

“Before BC” = 43% culls

“After BC” = 0.3% culls, 99% reduction

Olive scale,

Parlatoria oleae,

in CA, after

parasitoid

introduction

1.You need to start observations in control plots

up to several years before natural enemy

releases begin

2. Since there may be strong climatic differences

between the “before” and the “after” years, it

is best to continue some “before” sites without

releases (partial geographic design)

Potential problems or limitations with

the “before and after” design

ASSESSING IMPACT OF

CLASSICAL BIOLOGICAL CONTROL AGENTS

METHOD #2. Geographically

separated release and control plots

Assessment of pest density at sites with

releases vs ones with no releases of the

agent to be evaluated

Insert picture of ck lifestages

from C drive

Effect of Chilocorus kuwanae on euonymus scale,

Control Site (one of 15)- no predator release

Plant died in fall

Insert picture of ck lifestages

from C drive

Predator build up

Scale decline

Effect of Chilocorus kuwanae on euonymus scale.

Release site– one of 15

Predator build up

Scale decline

Insert picture of ck lifestages

from C drive

Predator build up

Effect of Chilocorus kuwanae on euonymus scale

(Unaspis euonymi), averaged over all research sites,

1991-1993 in New England

Predator present

Predator absent

Conclusion: sites with increasing scale typically

lacked the predator, while sites with the predator

typically decreased in scale density

Summary

Geographical plots to measure effect of phorid fly on fire

ants and effect of rapid range expansion on exp. design

Range expansion

post release of P.

tricuspis in two

years (black 1999,

dark grey 2000, and

light grey, 2001)

As a consequence of rapid expansion of fly after

release, control plots had to be relocated to somewhat

distant country (Madison, upper left

1. You need to define a pool of plots and assign

control or release treatments to them AT

RANDOM (this is often overlooked)

2. Control plots are sometimes invaded by the

natural enemy as it disperses. To avoid this,

greater separation must be used. This may

cause control plots to enter new ecological

zones or if control plots are grouped to

enhance separation from release plots, this

conflicts with random assignment.

Potential problems or limitations with the

“geographically separated plots” design

ASSESSING IMPACT OF

CLASSICAL BIOLOGICAL CONTROL AGENTS

METHOD #3. Cage Exclusion

Assessment of pest density inside and

outside cages excluding natural enemy

3 treatments-open cage, closed cage, no cage

Cage exclusion to measure pest density with

and without the key natural enemy

Closed cage

initially

impregnated

with DDT to

kill off any

pre-existing

parasitoids

in scales

Open cage is

intended to

allow full

parasitoids

full access to

scales

Uncaged

branch is a

check on

cage effects

Cage exclusion and CA red scale

Leaf clip cages are

used to assess

impacts of

parasitoids on

sedentary species

such as scale,

mealybugs and

whiteflies

Pest mortality inside and outside leaf clip cages

Example #1: Effect of parasites on CA red scale on ivy

Cage exclusion example #2: control of cassava mealybug

(Phenococcus manihoti) in Africa by Epidinocarsis lopezi

Cage exclusion example #2: control of cassava mealybug

(Phenococcus manihoti) in Africa by Epidinocarsis lopezi

Note log scale

>97% pest

reduction

1. Temperature may be increased inside cages,

causing pests to increase faster, perhaps

raising pest population density.

2. Humidity may be increased inside cages,

leading to higher rates of mortality from

fungal pathogens.

3. Pest progeny will be confined inside cages,

raising pest density by restricting pest

dispersal

Potential problems or limitations with

the “cage exclusion” design

ASSESSING IMPACT OF

CLASSICAL BIOLOGICAL CONTROL AGENTS

METHOD #4. Pesticide Exclusion

Assessment of pest density in plots

sprayed with a pesticide that kills the

natural enemy but not the pest vs.

unsprayed areas with the natural enemy

Use of pesticides to exclude natural enemy being evaluated

Pesticides must

1. Kill the natural enemy

2. Be safe to the pest

3. Not stimulate pest reproduction

Use of DDT

to exclude

natural

enemies of

California

red scale

Example #1 of chemical exclusion

70-fold pest

increase

Example #2 of chemical exclusion

Use of carbaryl to exclude predators of

Pacific mite in California vineyards

>30-fold pest

increase

Example #3: Use of pesticides to exclude

natural enemies of cassava mealybug

The residues of some pesticides stimulate the population

growth of such groups as mites or

1.2 eggs

for

controls

One of the problems in chemical exclusion: hormolygosis

1.6 eggs for low

residue treatment

0.6 eggs

for high

residue

treatment

Melaleuca invasion of the Everglades NP in the US-

converted miles of everglades to swamp forests

Thick Stands of Melaleuca in Florida

The melaleuca weevil

In FL in an insecticide exclusion trial, plots from which BC

agents were excluded showed lowered rates of tree mortality

than ones in which they were present

In FL in an insecticide exclusion trial, plots from which bc

agents were excluded showed stable tree populations in

contrast to declining tree numbers were insects were present

Insecticide exclusion trial on melaleuca in Florida

BC agents killed with insecticides unsprayed

1. It may be difficult to find a pesticide that has

no effect on the pest but kills the natural

enemy (easy if pest is a weed).

2. Pesticides are likely to eliminate the whole

natural enemy complex, not just the newly

introduced species.

3. Some pesticides stimulate pest fecundity,

raising pest density.

Potential problems or limitations with

the “pesticide exclusion” design

Ants that tend honey-dew-producing Homopera can

exclude natural enemies in some cases, acting as a

“biological check”

ASSESSING IMPACT OF CLASSICAL BIOLOGICAL

CONTROL AGENTS

Using life tables

Life tables

1. Provide an organizational framework for

data collection and comparison

2. Allow contributions of specific mortalities in

restricting population growth to be assessed

3. Allow separate effects of contemporaneous

mortalities to be distinguished

4. Series allow detection of density dependent

action of sources of mortality

Types of designs using life tables

1. Single- is a minimal description of what

happened in one generation

2. Series- a long series of life tables for one

population over many generations can be

used to look for mortality factors that act in

a density dependent way

3. Paired- pairs of life tables, for populations

with and without some feature of interest

(such as a new parasitoid) allows the impact

of the factor to be assessed

Cohort vs population based life tables

1. Cohort life table- is based on a group of

individuals that are created for the purpose

of tracking their fate. They are synchronized

and do not suffer exactly the same fates as

the real population since that may be more

protracted over time

2. Real population life table- based on samples

drawn from the population of interest in

some representative way

Cohort lifetables of

sessile species such

as leafminers are

readily obtained, as

a group can be

marked and easily

reencountered later

Mines of apple blotch leafminer (Phyllonorycter crataegella,

Lepidoptera: Gracillariidae) on apple

Sessile vs mobile stages or species

Cohort of immatures of apple blotch leafminer

(Phyllonorycter crataegella, Lepidoptera: Gracillariidae)

Mines retain a story of what happened to the pest, which can be

sorted out by what life stages, cast skins or residues are found

sap larva-pest

pupa-pest

pest pupal skins outside mines

larva of parasitoid

cadaver of

pest larva

Sample life table

How much mortality is enough

for pest control?

R0 is the best measure

Paired lifetable-example #1- apple blotch leafminers

Unsprayed

plot

Lifetable apple blotch leafminer (Phyllonorycter

crataegella) on unsprayed trees in Buckland, MA, 1981

65% larval parasitism

Popl’n increased by 1.8X

Manipulation of previous life table for apple blotch

leafminer (Phyllonorycter crataegella) (unsprayed trees in

Buckland, MA, 1981) with parasitism omitted

Unsprayed

Plot, parasitism omitted

Popl’n increased 8.6X

Real field lifetable for apple blotch leafminer

(Phyllonorycter crataegella) on sprayed trees (leafminer

resistant but parasitoids not resistant)

Sprayed Plot, parasitism

eliminated by pesticides

Popl’n increased 8.5X

Contrasting the Ro values for populations with and

with the natural enemy of interest in paired life tables

Ro values

• Unsprayed wild orchard 1.8X

• Table #1, parasitism removed 8.6X

• Sprayed orchard-

chemical check 8.5X

Conclusion: prediction made based on

lifetable from wild orchard matches

actual pest population growth in

sprayed orchard

Paired lifetable-example #2-whiteflies on poinsettia

B. tabaci, B strain adult and nymphs

Eretmocerus eremicus, parasitoid of the whitefly Bemisia

tabaci used in poinsettia via augmentative BC

Paired cohort lifetables to measure impact of Eretmocerus

eremicus on Bemisia tabaci in greenhouse poinsettia

Defined leaf areas are

repeatedly photographed

and survival of individual

whitefly nymphs recorded

Whitefly nymph

Life table for the whitefly Bemisia tabaci on poinsettia in

the absence of parasitoid releases

75% egg-to-adult-survival, R0 = 67 (90 eggs/F)

Life table for the whitefly Bemisia tabaci on poinsettia with

Eretmocerus eremicus releases

8% egg-to-adult-survival, R0 = 7.2 (90 eggs/F)

Parasitoids from the native

range imported to the USA

Four parasitoids

a)Oobius agrili (CH)

b)Tetrastichus

planipennisi (CH)

c)Spathius agrili (CH)

d)Spathius n sp. (Russia)

Major releases in USA

of first 3 began in 2008

Antanycolus spp. – a group of native parasitoids that now

dominate the EAB larval parasitoid guild

EAB: 2008-2009

An exploding pest population

Stage lx

dx di Factors

Egg 144 39 37 Infertility/predation

2 Oobius agrili

1st - 2nd 105 16 16 Tree resistance

3rd – 4th 89 4 2 Undetermined-Di/Com/other

1 T. planipennisi

1 Atanycolus sp

JL-PP 85 10 9 WP predation

1 Undetermined-Di/Com/other

(22) (22) (Overwintering WP-29%)

(Adults) (53) (3) (3) (Fungus disease -5%)

(F1 Eggs) 2529 (1:1 sex ratio/100 eggs/F)

R0 17.5

EAB: 2008-2009

Reduced to an expanding, not exploding, population

Stage lx

dx di Factors

Egg 417 78 78 Infertility/predation

0 Oobius agrili

1st - 2nd 339 55 54 Tree resistance

1 Undetermined-Di/Com/other

3rd – 4th 284 137 73 Undetermined-Di/Com/other

60 Atanycolus sp

4 T.planipennis/others

JL-PP 147 104 103 WP predation

1 Undetermined-Di/Com/other

(12) (12) (Overwintering WP-29%)

(Adults) (31) (2) (2) (Fungus disease -5%)

(F1 Eggs) 1450 (1:1 sex ratio/100 eggs/F)

R0 3.5

Comparison of EAB Population Growth in North America vs.

Asia

2009

MI -

gre

en a

sh

2010

MI -

gre

en a

sh

Asia

- gre

en a

sh

Asia-

Orie

ntal a

sh

Net P

opu

lation

Gro

wth

(R

0)

0

2

4

6

8

10

12

14

16

18

20

No woodpeckers60% egg parasitism60% larval parasitismLimited tree restance

Woodpecker predationDisease/competition/otherslittle egg parasitism~20% larval parasitismlimited tree resistance

High resistance (70%)

Use of life table data to determine if a source of mortality acts

in a density dependent manner

Rates of mortality to winter moth

from the tachind parasitoid Cyzenis

albicans over 6 years. Note upward

change in % from 1982-84

In principle, compensatory change in mortality by subsequent density dependent

factors can nullify effects of newly imposed mortality, as for example a release of

an egg parasitoid (e.g., Trichogramma releases)

Eggs Larvae Pupae Ro (given fertility

of 20 per F)

Natural popl’n In-100

kill-5%

Out-95

In-95

kill-50%

Out-48

In-48

% F-50

F Out-24

480 eggs

Ro 4.8

Trichogramma

releases added

In-100

kill-50%+5%

Out-45

In-45

kill-20%

Out-36

In-36

% F-50

F Out-18

360 eggs

Ro 3.6

Consequences Increases egg

mortality 10-fold

(5 to 55%)

Reduces

damaging

stage

(larvae by

53%)

But reduced new

year’s popl’n only

by 25%

Construction of life tables

How does one build accurate lifetables?

Previous life table examples were based on cohorts (a

marked set of individuals), not samples from a population

If insects are not sessile, you usually can’t follow the fate of

a set of individuals (a cohort). Rather, you have to estimate

the current numbers of each life stage by sampling

If insects are not sessile, you usually can’t follow the fate of

a set of individuals (a cohort). Rather, you have to estimate

the current numbers of each life stage by sampling

Change in density over time of Heliothis spp.

eggs and larvae

Density vs. numbers entering a stage

Components of a Lifetable

Data Used to Construct table

• Number entering each life stage (lx)

• Number dying, for each cause, in each stage (dx)

Calculations made from data

• Apparent mortality (qx) = dxi/lxi for a factor and

stage

• Irreplaceable mortality- dx/lx (stage 1)

• Marginal rate of mortality- rate of death for each

factor modified (when there are 2 or more sources

of mortality acting together) to reflect underlying

attack rates

Life tables are not built on density date. They ask for

numbers (summed over a generation) that enter each life

stage and how many die in each stage

In a generation, how many Colorado potato beetle eggs get laid?

Density and Recruitment

Estimating CPB egg

density and recruitment in

a potato field- stakes mark

sampling locations

Measuring lx for a stage

How can we measure

the numbers of eggs per

plant that get laid by a

whole generation of

Colorado potato

beetles?

Double Sample Method

1. Density– select sample plants and count egg

masses per plant on each date

2. Remove all egg masses found on plants sampled

to estimate density

3. Recruitment. On next sample date (3 or 4 days

later), recheck old density plants. Any egg

masses found on these plants had to have been

laid during the period between sample date.

This is a recruitment value (# eggs laid per plant

per time period).

Comparison of egg density over a series of dates and

numbers of new eggs laid per day in each time period

(recruitment to the stage)

Year 1

Year 2

Density and Recruitment

Blow up of

part of data

on previous

slide

Year 2, 1st

Generation

Year 2, 2nd

Generation

Peak density 275 (38%

of correct

number)

400 (25%)

Recruitment

(number entering

the stage)

729 1603

Relationship between peak density and number

of eggs laid per generation (recruitment) for

Colorado potato beetle

Measuring dx for a stage

How can we measure the

numbers of eggs per plant

that get parasitized by a

parasitoid species over

the course of a whole pest

generation?

Measuring the

host’s dx from

parasitism is

the same as

measuring the

lx to the

lifestage of the

immature

parasitoid

Host

popl’n

Imm. Par.

popl’n

Host gain/day

Imm par.

gain/day

Loss to next

host stage

Loss to parasite

emergence from host

Loss to death

Loss to death

Edovum puttleri, a eulophid, new association egg parasitoid

or Colorado potato beetle from Colombia

Measuring losses to parasitism

(dx parasitism from E. puttleri for Colorado potato beetle)

To measure total losses to parasitism:

1. Measure host lx to the egg stage as above

2. Measure semiweekly parasitoid attacks by

rearing eggs recruited in each time interval to

detect newly parasitized eggs

3. Sum all parasitized eggs in recruitment samples

for the whole egg generation (this is both

parasitoid lx to immature stage and dx from

parasitism for the host) (assumes eggs not

vulnerable to parasitism after one time period)

4. Generational loss to parasitism (qx of lifetable)

is dx parasitism/lx CPB egg \

Numbers of Colorado potato beetles laid per day and

parasitized per day over a whole generation = total loss (dx)

to egg parasitism

Cotesia glomerata is an imported parasitoid of Pieris rapae,

a pest of cabbage

Measuring losses to parasitism example #2

(dx parasitism from C. glomerata for P. rapae larvae)

Pieris rapae larvae parasitized by Cotesia glomerata, plus

cocoons from another previously parasitized larva

Measuring losses to parasitism #2

(dx parasitism from C. glomerata for P. rapae larvae)

1. Measure host lx to 1st larval instar (relatively

non mobile between leaves) using the double

sample methods described for CPB example)

2. Dissect all larvae (all instars) found on plants

sampled to measure P. rapae density.

3. Among dissected larvae (step 2), count only

those with small parasitoid eggs (= young

enough to be laid between sample dates).

4. Over whole insect generation, sum larvae with

small parasitoid eggs (step 3) and divide by

total host larvae (step1)

Summation for a generation of the total P. rapae larvae

entering the larval stage and the total number dying from

C. glomerata parasitism

Can good life table lx values be extracted from

samples of an insect stage’s density?

Only the graphical method of Southwood has ever been

widely used to obtain lx values from density sample data

Big issue: as

mortality in the

stage increases,

the value

obtained

becomes

progressively

TOO SMALL

Separating Simultaneous

Mortalities

Recovering rates of attack from observations on rates of

death, when two mortality agents act simultaneously

The fate of this

group is the critical

factor

Method is called “marginal rate of mortality”

When two

factors act

together, the

discrepancy

between

apparent

mortality and

the marginal

rate is

important only

when both

attack rates are

large

WHEN DEATH LEAVES NO TRACE: COUNTING LOSSES WHEN CADAVERS ARE MISSING

OR NOT DISTINCTIVE

1. Predation on non-attached prey stages

2. Host feeding by parasitoids

Predacious Hemiptera attacking a caterpillar

When the dead

caterpillar falls to

the ground, there is

nothing to mark

that it ever existed

Early option: ELISA

Current methods:

PCR-for prey DNA

An alternative is to seek remains of prey

inside predators

Caterpillar of apple blotch leafminer (Phyllonorycter

crataegella) that was killed by a host feeding parasitoid

Blood binding neck

of caterpillar to

skin of mine

Host fed

individuals are

findable in field

samples only if

cadavers are

stuck to plant

or contained in

leafmine, gall,

etc