**2. Challenges faced in breeding of Australian cultivars**

Breeding challenges have largely been dictated by unique biotic and abiotic limitations on production and the need for Australia to remain competitive with other producing countries that have lower labour costs. There has therefore always been a high focus on improving yield potential as a first priority and incremental increases in fibre quality to meet changing spinning requirements that might restrict the marketability of Australian cotton. Overlaid on these efforts has been targeted breeding for yield protection from losses caused from individual biotic and abiotic factors. Some of the early emphasis on breeding for host-plant resistance to insects has changed with the advent of genetically modified traits, but this has just moved the priorities to other challenges. A key asset to the breeding has been access to a diverse collection of germplasm from around the world, building on a quite diverse suite of base cultivars developed in the early 1970s and 1980s that have allowed CSIRO to tackle challenging insect, disease and weed issues.

resistance in *H. armigera*, although a pre-emptive *Bt* resistance management strategy was

Australian Cotton Germplasm Resources http://dx.doi.org/10.5772/58414 5

Over the past 10 years, silverleaf whitefly has occasionally reached pest status in some regions of the Australian cotton industry. It was first reported in cotton regions in 1994 [14], and the first major outbreak occurred in the Emerald production region in 2001/02. There are limited effective chemical control options for whitefly and they are also very expensive, so there is large emphasis on sampling protocols, thresholds and cultural control of this pest [15].

A range of other minor pests have occurred since 2000, many of which are related to the introduction and use of Bollgard II®. Reduced spraying for *Helicoverpa* has allowed survival of pale cotton stainers (*Dysdercus sidae*), jassids (*Austroasca viridigrisea*) and thrips (*Frankliniella schultzei* and *F. occidentalis*) late in the season and in some years these species have all caused

In Australia, weeds are generally less of a problem in cotton than insects such as *Helicoverpa*, but many areas do have a reasonably high incidence of problematic weeds particularly on former grazing and/or flood-prone land [16]. Some of the problem weeds are *Ipomea spp., Cyperus rotundas, Polymeria spp., Conyze bonariensis,* and *Datura spp.* Weed control systems have changed in the last decade: production has changed from a cultivation-based system with residual herbicides and hand hoeing, to a system of minimal cultivation, with the use of herbicide tolerant cultivars (glyphosate and glufosinate) and few if any residual herbicides [17]. This has caused weed composition to shift to glyphosate tolerant species (particularly *Conyze sp.*) and has the real risk of developing glyphosate resistant weed populations.

Many diseases of cotton are present in Australia and have necessitated the establishment of

Bacterial blight of cotton caused by *Xanthomonas axonopodis* pv. *malvacearum* was historically the most important and widespread disease in Australian cotton. This disease was present in all crops and averaged around 20% of bolls affected though the mid-to late-1980s. The use of standard differential cultivars showed that Race 18 of the pathogen was predominant [18] and that infected seed was the major factor in the epidemiology of the disease. The disease was partially controlled through the production of blight-free planting seed. The CSIRO breeding program has released cultivars with immunity to blight and this disease is no longer found in

Verticillium wilt caused by *Verticillium dahliae* has been an important disease in Australia for many years, especially in the older growing regions of the Namoi, Gwydir and McIntyre. It is assumed that this pathogen is endemic to Australia. To date, all the Australian Verticillium isolates that have been tested are categorised as 'non-defoliating' strains [19]. Despite not having the more severe 'defoliating' strains that occur in other countries, yield reductions of

Australian production systems due to 100% uptake of blight-resistant cultivars.

initiated [12].

**2.2. Weeds**

**2.3. Diseases**

damage and sporadically required control [12].

breeding projects to confer host resistance.

#### **2.1. Insects**

A number of important pests of cotton are found in Australia that have necessitated the establishment of intensive pest management programs [8]. Cotton breeders have initiated programs to address host plant resistance and eventually released local cultivars containing *Bt* traits to confer resistance to the most important *Lepidopteron* pests.

The key pests of cotton in Australia are the larvae of two lepidopteron species, *Helicoverpa armigera* and *H. punctigera*, together with cotton aphid (*Aphis gossypii*), green mirid (*Creontiades dilutus*), spider mites (*Tetranychus urticae*) and silverleaf whitefly (*Bemisia tabaci* B-Biotype). From the 1960's to early 2000's, pest management was highly reliant on use of insecticides, mostly broad spectrum organophosphates, carbamates and pyrethroids as well as endosulfan. Crops were sprayed around 12-16 times per season [9]. This has brought with it predictable problems of pesticide resistance, destruction of natural enemy populations resulting in pest resurgence and outbreaks of secondary pests such as aphids [10] and spider mites [11], health concerns and off-farm movement into sensitive riverine environments [9].

Deployment of Ingard® (Monsanto's Bollgard) cotton in 1996 reduced insecticide use on those crops by about 50%, but efficacy was limited due to declining expression of *Cry1Ac* through the growing season. Ingard® cotton was always seen as an interim technology and during this period its area was capped at 30% to reduce the risk that *Helicoverpa spp*. would develop resistance to this critical insecticidal protein. This regulation limited the influence of the technology on the industry. The strong reliance on insecticides continued on the remaining 70% of conventional cotton and led to ongoing selection for resistance to insecticides in *H. armigera*, secondary pest problems and selection of pesticide resistance in these secondary pests. For instance, by the early 2000's spider mites were resistant to organophosphates, the pyrethroid bifenthrin and chlorfenapyr, and cotton aphids were resistant to organophos‐ phates, the carbamate pirimicarb and pyrethroids [12].

The deployment of Bollgard II® in 2003/04, for improved resistance management and removal of the 30% area cap, resulted in a dramatic uptake of this technology. It also led to a massive decline in pesticide use (by about 85% [13]), especially relative to earlier years but also in comparison with contemporary conventional cotton that had also significantly reduced its reliance on pesticides. This technology essentially saved the cotton industry from insecticide resistance in *H. armigera*, although a pre-emptive *Bt* resistance management strategy was initiated [12].

Over the past 10 years, silverleaf whitefly has occasionally reached pest status in some regions of the Australian cotton industry. It was first reported in cotton regions in 1994 [14], and the first major outbreak occurred in the Emerald production region in 2001/02. There are limited effective chemical control options for whitefly and they are also very expensive, so there is large emphasis on sampling protocols, thresholds and cultural control of this pest [15].

A range of other minor pests have occurred since 2000, many of which are related to the introduction and use of Bollgard II®. Reduced spraying for *Helicoverpa* has allowed survival of pale cotton stainers (*Dysdercus sidae*), jassids (*Austroasca viridigrisea*) and thrips (*Frankliniella schultzei* and *F. occidentalis*) late in the season and in some years these species have all caused damage and sporadically required control [12].

#### **2.2. Weeds**

that have lower labour costs. There has therefore always been a high focus on improving yield potential as a first priority and incremental increases in fibre quality to meet changing spinning requirements that might restrict the marketability of Australian cotton. Overlaid on these efforts has been targeted breeding for yield protection from losses caused from individual biotic and abiotic factors. Some of the early emphasis on breeding for host-plant resistance to insects has changed with the advent of genetically modified traits, but this has just moved the priorities to other challenges. A key asset to the breeding has been access to a diverse collection of germplasm from around the world, building on a quite diverse suite of base cultivars developed in the early 1970s and 1980s that have allowed CSIRO to tackle challenging insect,

A number of important pests of cotton are found in Australia that have necessitated the establishment of intensive pest management programs [8]. Cotton breeders have initiated programs to address host plant resistance and eventually released local cultivars containing

The key pests of cotton in Australia are the larvae of two lepidopteron species, *Helicoverpa armigera* and *H. punctigera*, together with cotton aphid (*Aphis gossypii*), green mirid (*Creontiades dilutus*), spider mites (*Tetranychus urticae*) and silverleaf whitefly (*Bemisia tabaci* B-Biotype). From the 1960's to early 2000's, pest management was highly reliant on use of insecticides, mostly broad spectrum organophosphates, carbamates and pyrethroids as well as endosulfan. Crops were sprayed around 12-16 times per season [9]. This has brought with it predictable problems of pesticide resistance, destruction of natural enemy populations resulting in pest resurgence and outbreaks of secondary pests such as aphids [10] and spider mites [11], health

Deployment of Ingard® (Monsanto's Bollgard) cotton in 1996 reduced insecticide use on those crops by about 50%, but efficacy was limited due to declining expression of *Cry1Ac* through the growing season. Ingard® cotton was always seen as an interim technology and during this period its area was capped at 30% to reduce the risk that *Helicoverpa spp*. would develop resistance to this critical insecticidal protein. This regulation limited the influence of the technology on the industry. The strong reliance on insecticides continued on the remaining 70% of conventional cotton and led to ongoing selection for resistance to insecticides in *H. armigera*, secondary pest problems and selection of pesticide resistance in these secondary pests. For instance, by the early 2000's spider mites were resistant to organophosphates, the pyrethroid bifenthrin and chlorfenapyr, and cotton aphids were resistant to organophos‐

The deployment of Bollgard II® in 2003/04, for improved resistance management and removal of the 30% area cap, resulted in a dramatic uptake of this technology. It also led to a massive decline in pesticide use (by about 85% [13]), especially relative to earlier years but also in comparison with contemporary conventional cotton that had also significantly reduced its reliance on pesticides. This technology essentially saved the cotton industry from insecticide

*Bt* traits to confer resistance to the most important *Lepidopteron* pests.

concerns and off-farm movement into sensitive riverine environments [9].

phates, the carbamate pirimicarb and pyrethroids [12].

disease and weed issues.

4 World Cotton Germplasm Resources

**2.1. Insects**

In Australia, weeds are generally less of a problem in cotton than insects such as *Helicoverpa*, but many areas do have a reasonably high incidence of problematic weeds particularly on former grazing and/or flood-prone land [16]. Some of the problem weeds are *Ipomea spp., Cyperus rotundas, Polymeria spp., Conyze bonariensis,* and *Datura spp.* Weed control systems have changed in the last decade: production has changed from a cultivation-based system with residual herbicides and hand hoeing, to a system of minimal cultivation, with the use of herbicide tolerant cultivars (glyphosate and glufosinate) and few if any residual herbicides [17]. This has caused weed composition to shift to glyphosate tolerant species (particularly *Conyze sp.*) and has the real risk of developing glyphosate resistant weed populations.

#### **2.3. Diseases**

Many diseases of cotton are present in Australia and have necessitated the establishment of breeding projects to confer host resistance.

Bacterial blight of cotton caused by *Xanthomonas axonopodis* pv. *malvacearum* was historically the most important and widespread disease in Australian cotton. This disease was present in all crops and averaged around 20% of bolls affected though the mid-to late-1980s. The use of standard differential cultivars showed that Race 18 of the pathogen was predominant [18] and that infected seed was the major factor in the epidemiology of the disease. The disease was partially controlled through the production of blight-free planting seed. The CSIRO breeding program has released cultivars with immunity to blight and this disease is no longer found in Australian production systems due to 100% uptake of blight-resistant cultivars.

Verticillium wilt caused by *Verticillium dahliae* has been an important disease in Australia for many years, especially in the older growing regions of the Namoi, Gwydir and McIntyre. It is assumed that this pathogen is endemic to Australia. To date, all the Australian Verticillium isolates that have been tested are categorised as 'non-defoliating' strains [19]. Despite not having the more severe 'defoliating' strains that occur in other countries, yield reductions of up to 20% have been recorded [20]. The majority of cultivars now grown commercially have relatively strong resistance to Verticillium wilt, but in seasonal conditions that favour the disease (temperature between 22-27°C [21]) and high number of rain days during the boll filling period [Allen S.J, pers comm.], significant productions losses still occur in some regions so improving resistance remains a breeding focus at a somewhat lower priority than other diseases.

Alternaria leaf spot caused by *Alternaria macrospora* has caused severe defoliation of cotton grown on less fertile soils when the crop is exposed to extended periods of wet weather. These conditions occasionally occur in some Queensland growing regions. All upland cultivars currently grown have some degree of resistance, so the disease is considered of minor

Australian Cotton Germplasm Resources http://dx.doi.org/10.5772/58414 7

Reniform nematodes, *Rotylenchulus reniformis*, associated with stunting of cotton plants have only recently been identified in Australia [32] and are yet to be considered of economic importance. There was an isolated detection in 2003 in central Qld, but a more widespread identification was made in 2012, so it may become more important with time and pre-emptive

A range of other industry challenges are also being, or need to be, addressed through breeding. There is a constant awareness of the dangers of having a narrow genetic base from which cultivars are developed. However, it is also acknowledged that broadening the genetic base through the introduction of exotic germplasm can often have a detrimental effect on yield [33]. In Australia, there is clear evidence of continuing yield improvement through selection within a 'narrowing' germplasm pool. In a high input, high yield industry such as Australia, there is a requirement for continued yield improvement to counter the ever increasing cost of pro‐ duction. While some of these improvements can come from improved management, historical data indicates that new cultivars are expected to provide at least 50% of the increase [2].

Many of the new challenges are related to the introduction of new technology or new envi‐ ronmental standards. The introduction of GM insect resistance traits into Australia in the mid 1990s simplified many aspects of pest management. However, it was soon discovered that the combination of high fruit retention of insect-protected crops together with early-maturity genotypes resulted in crops that had a fruiting cycle shorter than the available growing season and thus were not capitalising on the inherent yield potential of the cultivars. This had a detrimental effect on yield, so longer season, more indeterminate germplasm was subsequent‐ ly used to deploy the GM traits. This remains a challenge as the industry expands into the southern (shorter season) regions of Australia where yield expectations of growers are similar to the fuller season traditional regions and may require a different approach to breeding.

Most Australian cotton is exported to Asia and spun on ring-spinning frames. Although this spinning technology has been used for around 100 years, modern machines continue to run faster and spinning companies demand greater throughput. This places greater stresses on the fibre, so there is a continual demand for higher quality fibre. There are well documented negative associations between yield and quality [34] which makes the simultaneous improve‐ ment of both yield and fibre quality difficult. In addition, breeders are also expected to address changing future fibre quality requirements due to changes in spinning technology or the yarn

At a time of increased awareness of the potential for climate change, greater scrutiny is also being placed on climate-related aspects of agricultural production [35]. Australia already has

importance [20].

breeding for resistance may be desirable.

that is required by the textile industry.

**2.4. Other industry challenges**

Fusarium wilt caused by *Fusarium oxysporum* f sp. *vasinfectum*, was first recognised in Australia in 1993 [22] and is characterised by causing plant mortality at any time throughout the season, from seedling emergence through to harvest. By the end of 1999, the disease was present in six of the ten cotton production areas in Eastern Australia, but only widespread in two of those areas [23]. Subsequently, it has been shown that this is a unique strain of Fusarium and is substantially different to strains that are pathogenic to cotton in other countries [24]. This new Australian strain is identified as race 6 and is associated with alkaline clays, absence of nematodes [25], seedling death and an optimum temperature of 18 to 23°C [26]. Fusarium in the USA are races 1 and 2 and are associated with acid sands and nematodes [25], symptoms occur in mid-season with no plant mortality and an optimum temperature of 30 to 32°C [27]. It is assumed that the Australian pathogen evolved locally on weeds within our production system [28]. The extreme virulence and persistence of this pathogen and its ability to be readily transported in soil, water or trash raised concern within the industry when it was first discovered [29], particularly because most commercially grown cultivars in 1994 were highly susceptible. Production losses of virtually 100% have been reported for some fields. Substantial progress has been made in developing more resistant cultivars. However, for fields that have a high level of inoculum, the most resistant cultivar may still only have 10% survival in seasons that favour the disease so it remains a significant breeding challenge.

Cotton bunchy top disease (CBT) was first recorded in the late 1990s when a severe outbreak occurred in several Australian growing regions. Economic losses in that season alone were estimated at AUD\$70 million [30], with nearly all Australian cotton cultivars being susceptible to the disease. The cotton aphid (*Aphis gossypii*) was identified as the vector and since then, there have only been sporadic outbreaks as the vector has generally been well controlled but its incidence is now very widespread across the industry. The causal agent has subsequently been identified as a Polerovirus and has some similarities to cotton blue disease found in Africa, Asia and the Americas [30]. Currently no commercial cultivars in Australia have resistance to this disease, although a readily accessible source of resistance has been identified in Australian material and cultivars are due for release around 2017.

Black root rot caused by *Thielaviopsis basicola* was first reported on cotton in Australia in 1990 [31]. It is now present in almost all cotton fields in NSW, as well as many in Qld. Diseased plants show stunted growth early in the season when the cooler temperatures favour the pathogen. While the pathogen does not kill plants, it can cause significant yield loss in cooler and shorter season regions due to the delay in crop maturation. The wide host range of the pathogen and persistence of the chlamydospores makes effective control through crop rotation very difficult. No *G. hirsutum* or *G. barbadense* cotton cultivars have been shown to have any significant resistance to this pathogen, so breeders must look to primary, secondary or tertiary germplasm or to GM approaches for control of this disease.

Alternaria leaf spot caused by *Alternaria macrospora* has caused severe defoliation of cotton grown on less fertile soils when the crop is exposed to extended periods of wet weather. These conditions occasionally occur in some Queensland growing regions. All upland cultivars currently grown have some degree of resistance, so the disease is considered of minor importance [20].

Reniform nematodes, *Rotylenchulus reniformis*, associated with stunting of cotton plants have only recently been identified in Australia [32] and are yet to be considered of economic importance. There was an isolated detection in 2003 in central Qld, but a more widespread identification was made in 2012, so it may become more important with time and pre-emptive breeding for resistance may be desirable.

#### **2.4. Other industry challenges**

up to 20% have been recorded [20]. The majority of cultivars now grown commercially have relatively strong resistance to Verticillium wilt, but in seasonal conditions that favour the disease (temperature between 22-27°C [21]) and high number of rain days during the boll filling period [Allen S.J, pers comm.], significant productions losses still occur in some regions so improving resistance remains a breeding focus at a somewhat lower priority than other

Fusarium wilt caused by *Fusarium oxysporum* f sp. *vasinfectum*, was first recognised in Australia in 1993 [22] and is characterised by causing plant mortality at any time throughout the season, from seedling emergence through to harvest. By the end of 1999, the disease was present in six of the ten cotton production areas in Eastern Australia, but only widespread in two of those areas [23]. Subsequently, it has been shown that this is a unique strain of Fusarium and is substantially different to strains that are pathogenic to cotton in other countries [24]. This new Australian strain is identified as race 6 and is associated with alkaline clays, absence of nematodes [25], seedling death and an optimum temperature of 18 to 23°C [26]. Fusarium in the USA are races 1 and 2 and are associated with acid sands and nematodes [25], symptoms occur in mid-season with no plant mortality and an optimum temperature of 30 to 32°C [27]. It is assumed that the Australian pathogen evolved locally on weeds within our production system [28]. The extreme virulence and persistence of this pathogen and its ability to be readily transported in soil, water or trash raised concern within the industry when it was first discovered [29], particularly because most commercially grown cultivars in 1994 were highly susceptible. Production losses of virtually 100% have been reported for some fields. Substantial progress has been made in developing more resistant cultivars. However, for fields that have a high level of inoculum, the most resistant cultivar may still only have 10% survival in seasons

Cotton bunchy top disease (CBT) was first recorded in the late 1990s when a severe outbreak occurred in several Australian growing regions. Economic losses in that season alone were estimated at AUD\$70 million [30], with nearly all Australian cotton cultivars being susceptible to the disease. The cotton aphid (*Aphis gossypii*) was identified as the vector and since then, there have only been sporadic outbreaks as the vector has generally been well controlled but its incidence is now very widespread across the industry. The causal agent has subsequently been identified as a Polerovirus and has some similarities to cotton blue disease found in Africa, Asia and the Americas [30]. Currently no commercial cultivars in Australia have resistance to this disease, although a readily accessible source of resistance has been identified in Australian

Black root rot caused by *Thielaviopsis basicola* was first reported on cotton in Australia in 1990 [31]. It is now present in almost all cotton fields in NSW, as well as many in Qld. Diseased plants show stunted growth early in the season when the cooler temperatures favour the pathogen. While the pathogen does not kill plants, it can cause significant yield loss in cooler and shorter season regions due to the delay in crop maturation. The wide host range of the pathogen and persistence of the chlamydospores makes effective control through crop rotation very difficult. No *G. hirsutum* or *G. barbadense* cotton cultivars have been shown to have any significant resistance to this pathogen, so breeders must look to primary, secondary or tertiary

that favour the disease so it remains a significant breeding challenge.

material and cultivars are due for release around 2017.

germplasm or to GM approaches for control of this disease.

diseases.

6 World Cotton Germplasm Resources

A range of other industry challenges are also being, or need to be, addressed through breeding. There is a constant awareness of the dangers of having a narrow genetic base from which cultivars are developed. However, it is also acknowledged that broadening the genetic base through the introduction of exotic germplasm can often have a detrimental effect on yield [33]. In Australia, there is clear evidence of continuing yield improvement through selection within a 'narrowing' germplasm pool. In a high input, high yield industry such as Australia, there is a requirement for continued yield improvement to counter the ever increasing cost of pro‐ duction. While some of these improvements can come from improved management, historical data indicates that new cultivars are expected to provide at least 50% of the increase [2].

Many of the new challenges are related to the introduction of new technology or new envi‐ ronmental standards. The introduction of GM insect resistance traits into Australia in the mid 1990s simplified many aspects of pest management. However, it was soon discovered that the combination of high fruit retention of insect-protected crops together with early-maturity genotypes resulted in crops that had a fruiting cycle shorter than the available growing season and thus were not capitalising on the inherent yield potential of the cultivars. This had a detrimental effect on yield, so longer season, more indeterminate germplasm was subsequent‐ ly used to deploy the GM traits. This remains a challenge as the industry expands into the southern (shorter season) regions of Australia where yield expectations of growers are similar to the fuller season traditional regions and may require a different approach to breeding.

Most Australian cotton is exported to Asia and spun on ring-spinning frames. Although this spinning technology has been used for around 100 years, modern machines continue to run faster and spinning companies demand greater throughput. This places greater stresses on the fibre, so there is a continual demand for higher quality fibre. There are well documented negative associations between yield and quality [34] which makes the simultaneous improve‐ ment of both yield and fibre quality difficult. In addition, breeders are also expected to address changing future fibre quality requirements due to changes in spinning technology or the yarn that is required by the textile industry.

At a time of increased awareness of the potential for climate change, greater scrutiny is also being placed on climate-related aspects of agricultural production [35]. Australia already has a limited water supply, so increases in water-use efficiency (WUE) are being sought through management and breeding. In addition, the energy consumed and the emissions produced by a cotton crop are also being closely examined. Due to the heavy influence that yield has in the WUE calculation, this must continue to be the major focus of breeding programs. The efficient use of other resources, such as nitrogen fertiliser, will significantly influence the emissions from a crop [36, 37], so breeding programs also have responsibilities to continue research in this area.

**Species Number of accessions Genome\***

Australian Cotton Germplasm Resources http://dx.doi.org/10.5772/58414 9

*G. tomentosum* 6 AD3 *G. mustelinum* 1 AD4 *G.darwinii* 1 AD5

*G. herbaceum* 39 A1 *G. arboreum* 211 A2 *G. anomalum* 4 B1 *G. logicalyx* 1 F1 *G. thurberi* 5 D1 *G. trilobum* 1 D8 *G. davidsonii* 2 D3-d *G. klotzschianum* 1 D3-k *G. armourianum* 1 D2-1 *G. harknessii* 1 D2-2 *G. aridum* 1 D4 *G. raimondii* 1 D5

*G.sturtianum* 64 C1 *G.robinsonii* 3 C2 *G. stocksii* 3 E1 *G. somalense* 2 E2 *G.australe* 159 G *G.nelsonii* 39 G *G.bickii* 28 G1 *G.costulatum* 1 K *G.cunninghamii* 2 K *G.enthyle* 1 K *G.exiguum* 3 K *G.nobile* 1 K *G.pilosum* 1 K *G.populifolium* 1 K *G.pulchellum* 1 K *G.rotundifolium* 3 K

**Table 2.** Number of wild primary, secondary and tertiary tetraploid and diploid accessions in Australian collections.

Primary gene pool

Secondary gene pool

Tertiary gene pool

unclassified 8

\* As per [39]
