USE OF BREED RESOURCES FOR POULTRY EGG AND MEAT PRODUCTION

R.W. FAIRFULL and R.S. GOWE*CANADA

SUMMARY

The preservation of poultry gene pools, and crossing strains or breeds with respect to heterosis, reciprocal effects, selection and prediction are discussed. Published values show significant heterosis for egg production, sexual maturity, body and egg weight, and viability. Reciprocal differences are typically lower. Heterosis is due to allelic (dominance or overdominance) or non-allelic (epistasis) gene interaction. Egg production and related traits have a major epistatic component. Reciprocal differences are due to sex-linked effects (Z-chromosome genes) or maternal effects (plus w-linked genes). Sex-linked effects are generally more important than maternal effects, but maternal effects are important for early growth, and effects on viability caused by disease organisms and egg transmitted antibodies. Heterosis and reciprocal effects are influenced by age and environment. Performance across a range of environments and ages is correlated, but rankings of crosses change. In broilers, specialized sire and dam lines are used to increase effective selection pressure. In layers, within-line selection and crossing is illustrated. Data from four Leghorn strains and their crosses was used to study prediction of cross performance. For sexual maturity and body weight, expectations from pures were correlated with 2-way, 3-way and F2 means. For egg production, expectations from 2-ways were correlated with 3-way and F2

means. Predictions of 4-way means were poor.

INTRODUCTION

This paper will discuss the evaluation and utilization of breed resources of poultry for egg and meat production under two topics: crossing strains or breeds and predicting cross performance: and preservation of strains orbreeds. Crossing strains or breeds will be emphasized.

In the poultry industry, the concept of breed has been almost eliminated. Breeders deal with numbered strains that started as White Leghorn or White Plymouth Rock or Dark Cornish Game or some other breed or synthesis of breeds, but to which genes for feathering rate, feather colour, disease resistance, dwarfing, etc. have been added as required from other "breeds". Thus, the populations no longer descend from "pure-bred" ancestors. Also, for many generations, these gene pools have been selected exclusively for traits of economic merit. Thus, the strains no longer meet the breeds' physical standards. The commercial breeder uses White Leghorn-like or Rhode Island Red-like strains for egg stocks, or White Rock-like meat dam strains and White Cornish-like meat sire strains, none of which are likely to be "pure-bred" or meet breed standards.

The poultry industry has a history of utilizing breed crosses and more recently, strain crosses in poultry production. One reason has been to take advantage of heterosis: in individuals for viability, egg production and size (egg-type birds), growth (meat-type birds) and efficiency of feed utilization: and in dams for viability, egg production and size, hatchability, and effi­ciency of feed utilization. Since the primary or grandparent stocks have to

"Animal Research Centre, Agriculture Canada, Ottawa, Ontario, Canada, K1A 0C6242

reproduce the parent stocks which in turn are used to produce the commercial egg or meat stocks, it is relatively easy to introduce crossing into the process. The development of a breeding program utilizing crossing also makes it possible for the breeder to combine stocks that compliment one another, and to develop specialized sire and dam lines - particularly in meat stocks. Crossing to produce the commercial stock also permits the commercial breeder to control the release of primary (pure) lines.

There has always been great interest in the prediction of performance. The components of cross performance are important in the design of breeding programs. The biological problem is to identify and interpret the pertinant effects. These effects then define models which can be examined empirically.

There are many reasons to preserve poultry gene pools. In poultry, gene transfer is close to reality. Preserved poultry gene pools are insurance against crisis or change, and are potential sources of genes for transfer. In addition, local strains or breeds provide the possibility of combining stock adapted to local conditions with high performance stocks by development of synthetic populations, crossing of local and high performance stocks or the transfer of adaptation genes to the performance line.

CROSSING STRAINS OR BREEDSHeterosis

Heterosis, the deviation between the cross mean and the mid-parent mean, is an important phenomenon in poultry. Heterosis has been studied extensively in layers: however, somewhat less experimental evidence exists for meat type fowl. In many crossing studies, the pure strains or reciprocals were not included and while these studies may yield information on combining ability, they do not contribute information on heterosis or reciprocal differences.

A survey of literature values for percentage heterosis (percent of mid­parent mean) of 2-way crosses is presented in Table 1. There is consistent and substantial heterosis for hen-housed egg production (HHP), total egg yield or mass (EYLD) and feed conversion (CONV). Heterosis for egg weight (EWT) is low but consistent. Heterosis for body weight (BWT) is close to zero at 1 week of age and earlier but increases to 2 to 5% by about 8 to 10 weeks and values for mature birds, fall in the same range. Heterosis for sexual maturity (AGFE) is typically -3 to -7%. Viability (VIAB) usually has positive heterosis below 5%, but negative and very high values are not uncommon. The egg quality traits: specific gravity (SGR, an indirect measure of eggshell strength), Haugh units (HAU, a measure of albumen quality) and percentage of eggs free of blood spots (BSP), are characterized by little or no heterosis. Thus, most of the major production and reproduction traits in egg and meat type poultry are influenced beneficially by heterosis.

There are few reports on the effects of crossing on fertility (FERT) and hatchability (HATC). Table 2 shows FERT and HATC from Fairfull et al.(1986). The hybrid embryo improves HATC (also shown by Merritt and Gowe, 1960: Friars et al., 1963: Morris and Binet, 1966) but has no effect on FERT (Table 2). The hybrid dam increases FERT (also shown by Horn, 1974), but a hybrid sire has no effect on FERT or HATC (Table 2).

With the virtual disappearance of pure breeds in commercial poultry, it is not possible to differentiate meaningfully between breed and strain crosses.

243

This was not the case 20 to 30 years ago when breed crosses exceeded crosses of strains (Dickerson, 1973). Although classical pure breeds are not used in broiler production, there are large differences between sire lines with a high proportion of Cornish genes and dam lines usually with a high proportion of White Rock genes. Such large physical differences are not seen in the strains used to produce eggs either brown or white. However, samples (strains, lines) of white egg ("Leghorn") genotypes or brown egg ("Rhode Island Red") genotypes can have very dissimilar performance and different crossing characteristics.

TABLE 1. Mean percentage heterosis calculated from published values9

Author11 Year Breedc VIAB HHP EYLD CONV AGFE BWTd EWT

Aggarwal 1979 C2&R2 0-2eBlyth 1970 L6 5 21-»30 -7->-9 lowCole 1973 L2 -4->l 10-»13 -3-»-4 3-»5m 0->3Fairfull 1986 L4 1 12 14 -3e -3 5» 2Flock 1980 L2 3 20 25 4M 4Friars 1963 T2 44 1-»4eGowe 1982a L6 1 6-»8 10 -6 -3 2->4m 2Gowe 1982af L6 l->7 13->16 16 -8 2->5m 2Hale 1965 L1&S1 5->15Horn 1980 R2 0-»5 8-»ll 4->6m 2->3Horn 1980f R2 0->l 6-»13 4->6m 3Hull 1963 Lll 0-»3 3->79 0-»-4 3->5m 0^1Jerome 1960 T3 -2-»16IMerritt 1960 Nllh 4->5 18-»20 -3-»-5 2-»6e 2->3Morris 1966 A1&L1 40 -7Nestor 1971 T3 2e-»0iPodger 1980 A2&L2 26 -14 5«Roy 1981 N3 0-»5eSheridan 1977 A1&L1 -l-*24 15-*30 21-»23 0-»-13 -3->-7 3-*5Taran 1972 H1&L1 0->3 -2->0m 0-»2Tijen 1977 L2 1 2Tijen 1977 R2 -1 1Tijen 1977 L2&R2 13 2Wearden 1965 L3 -9 -3 0MWearden 1965 R3 0 11 0MWearden 1965 L3&R3 -2 14 0MWilleke 1982 E2 -2 4 5 -3 3M 0Yoo 1984 A2&L2 29 31 -16 -4 5

aMean heterosis calculated for 2-way crosses (F3). bFirst author only. cA=Australorp, B=Peachblows, C=Cornish, E=Egg type, H=New Hampshire,L=Leghorn, N=Meat type, P=Plymouth Rock, R=Rhode Island Red, S=Synthetic, T=Turkey. A number after the breed indicates the number of lines or strains. dFor body weight, the superscript: E=early body weight in weeks 1 to 10; I=intermediate body weight in weeks 11 to 20: M=body weight around or after sexual maturity. determined over a four week period after sexual maturity. ^Second cycle of egg production following an induced moult. 9Egg production of surviving hens. bBl, H2, P4, Rl, S2 & Wl.

Two genetic mechanisms are responsible for heterosis: interactions between alleles known as dominance of which overdominance is a special case: and inter­actions between non-allelic genes known as epistasis. In livestock species, dominance has generally been believed to be the most important or the only im-

244

portant mechanism of heterosis (Pirchner, 1969: Sheridan, 1981). it has been known for some time that epistasis may play a role in heterosis in poultry (Dickerson, 1965). However, until recently this was largely ignored. In both strain and breed crosses of layers, significant epistasis has been found for egg production traits (Sheridan and Randall, 1977; Fairfull et al. , 1986). In Australorp x Leghorn crosses, Sheridan and Randall (1977) found that heterosis for feed conversion and egg production was due almost entirely to epistasis. Results from Fairfull et_ al. (1986) for Leghorn strain crosses are shown in Table 3. Epistasis was a major component of heterosis for egg production; however, heterosis for body weight, egg weight and sexual maturity appeared to be largely due to dominance (Fairfull et al., 1986). The traits in which heterosis had the greatest magnitude were those influenced by epistasis.

TABLE 2. Fertility and hatchabil- ity of eggs of 4 unrelated Leghorn pure strains and their crosses

TABLE 3. Mean performance of 4 Leghorn strains and their crosses housed 2 and 3 hens/cage (from Fairfull et al., 1986)

Embryo Eggs Set FERT HATC Type Housed VIABa HHPa AGFEb EWTC

Pure strain 4194 87.5 81.8 Pure strain 1024 85.7 232 147 56.52-strain (F^) 11361 86.4 89.0 2-strain 1632 87.6 260 143 57.83-straina 10374 91.1 91.2 3-strain 2304 87.5 255 143 57.84-strain 10279 90.1 91.5 4-strain 2304 87.0 247 143 57.7f 2 7001 92.4 89.7 *2 1632 87.2 247 147 57.1

aDam is a 2-strain cross. a134 to 497d. bDays to 50% egg prod. c240d.

Heterosis for viability is very complex and warrants separate comment. Crosses do not always exhibit heterosis for viability. Heterosis for viability may result from a complex interaction of specific effects for several diseases and stresses. The highly variable results for viability are likely due to the effects of different pathogens. Thus, the viability and morbidity resulting from each disease should be viewed as a separate trait. Unless test conditions are defined in terms of the pathogens present and their virulence, overall via­bility is not a good trait to illucidate heterosis or other genetic mechanisms. Under specific pathogen-free conditions there is little heterosis and so it is likely that heterosis for most traits is affected by pathogens (J.S. Gavora, personal communication).

Reciprocal effects

Reciprocal effects are due largely to sex-linked (Z-chromosome genes) and maternal (confounded with w-linked genes) effects. Sex-linkage likely accounts for most of these effects since Z-chromosome genes are known to affect egg production and other traits in the chicken. Except for early growth rate (Merritt and Gowe, 1965; Al-Murrani, 1978: Aggarwal et al., 1979), and effects of disease organisms, maternal effects are not generally considered important (Wearden et al., 1965: Bernon and Chambers, 1985). There are exceptions. Lowe and Garwood (1981) suspected that the w-chromosome was responsible for maternal effects for body weight and egg weight, and Willeke (1982) found maternal effects for egg weight.

Recent reports associating endogenous viral (ev) genes with slow feathering (K) genes on the Z-chromosome illustrate one mechanism of sex-linked effects. Harris et al. (1984) reported a high rate of horizontal lymphyoid leukosis virus (LLV) infection in slow feathering (K) birds. The presence of ev genes

245

impairs immune response to exogenous LLV (Crittenden et al., 1984), and Bacon et al. (1985) reported a close association between the K gene and the endoge­nous LLV gene ev-21. This may explain the higher rates of horizontal infection in slow feathering birds and their progeny.

Resistance to disease and parasite infection appear to be instances where maternal effects are more important than sex-linkage. Reciprocal differences in viability observed after challenge with Eimeria s£. (Jeffers et al., 1970) and Marek's disease (Krosigk et al., 1972) were ascribed to maternal effects since they were present in male as well as female progeny. The presence or absence of maternal antibody can contribute to a maternal effect in the case of viability (Nordskog and Pevzner, 1977). Offspring without maternal antibody protection are vulnerable to disease before they develop immune competence. As some strains or breeds have no genetic resistance to certain diseases and susceptibility is often dominant, chicks from a susceptible male by resistant female cross will have neither maternal antibody nor resistance. Chicks from the opposite cross female (susceptible) that was exposed to the disease and built up a high antibody titre would pass on a high level of passive antibodies to their progeny. This in turn could protect the chicks and prevent antibody build up so that the next generation would have little or no protection.

Congenitally transmitted diseases may also contribute to maternal effects. Production characters are depressed more in hens congenitally infected with LLV than in horizontally infected hens (Payne et al,., 1982: Fadly and Okazaki, 1982: Fairfull et al,., 1984). Reciprocal differences in crosses would depend on the rate of congenital infection. The rate of infection in 2-way crosses increases over that of the parental strains (Fairfull et al., 1984) so that this type of reciprocal effect could be amplified in 3-way and higher crosses due to relaxed selection (Gavora and Spencer, 1985).

Reciprocal differences as a percentage of the mid-parent mean are presented in Table 4. Reciprocal differences are generally smaller than heterosis values, but substantial for most traits. Although reciprocal differences for specific gravity, Haugh units and blood spots percentage are low, they are more important for these traits than heterosis.

Environment

Environment affects the magnitude of heterosis (Barlow, 1981). In poultry, changes in heterosis have been found with respect to nutritional treatment (Hull <et al., 1963: Emsley et al., 1980) and housing (Blyth and Sang, 1960: Horn et al.. 1980: Gowe and Fairfull, 1982a). Two recent studies are of particular interest because: their results are remarkably consistent: relatively high performance stocks were used: and the hens were molted to obtain data on a second cycle of production. One study used Rhode Island Red strain crosses housed 2 and 4 hens/cage (Horn et al., 1980) and the other White Leghorn strain crosses housed 1 and 3 hens/cage (Gowe and Fairfull, 1982a). In both cases, as the stocking rate increased, heterosis for the egg production traits increased, heterosis for sexual maturity and egg weight remained constant, and heterosis for body weight declined. In general, heterosis for egg production increased with the severity of the environment.

Environment also affects reciprocal differences (Gowe and Fairfull, 1982a). Reciprocal differences at the higher stocking rate (3 hens/cage) were equal to or greater than those at the lower stocking rate (1 hen/cage) and reciprocal differences for individual pairs of reciprocal crosses varied significantly

246

between environments.

TABLE 4. Mean reciprocal difference as a percentage of the mid-parent value3

Authorb,c Year VIAB HHP EYLD CONV AGFE: BWTd EWT SGR HAU BSP

Blyth 1960 8-»9 7-»8Cole 1973 l->9 6-»19 l->3 1->4m 1Fairfull 1986 2 4 4 2 3 2 1 1 2Friars 1963 47 1->5eGowe 1982a 4->8 3-»4 3 1 2->3 2->4 2 1 2 2Gowe 1982ae1 7 7->8 6 4 3 3 2->3 2 3-»5Morris 1966 9 2Nestor 1971 lO5- ^ 1Roy 1981 4->21eSheridan 1977 2->15 3->10 2->4 4-»8 0->l 5Taran 1972 0->2 0-*5m 0-»2Tijen 1977 1 1 1Tijen 1977 1 1 1Tijen 1977 1 2 1Wearden 1967 14 6 3 0MWilleke 1982 2 1 3 1 2m 2

Calculated for 2-way crosses (Ft ) bFirst author only. cSee note cin Table 1. dSee note d in Table 1. eSecond cycle of egg production.

Correlations between mean performance of hens housed 1/cage and that of hens housed 3/cage for 24 Leghorn strain crosses in two cycles of production for laying viability (VIAB), HHP and EWT are shown in Table 5. These values reflect deviations in performance due to heterosis and reciprocal differences as well as sampling error. Performance with 1 hen/cage is highly correlated with that of 3/cage for HHP and EWT especially in the first production cycle. The values for VIAB are lower. These correlations indicate that the better crosses with 1 hen/cage tend to be the better crosses with 3/cage, but the correlations are far from perfect especially for VIAB so that the overall rank of the crosses in the two environments change. Horn (1982) calculated genetic correlations from sire half-sibs housed 2 and 4 hens/cage. For HHP, he obtained correlations of 0.69 in the 1st cycle and 0.72 in the 2nd cycle of egg production between the two environments.

Heterosis for some traits is affected by age: during early growth (Aggarwal et al■, 1979: Roy and Kumar, 1981): and over production cycles (Horn et̂ al■, 1980: Gowe and Fairfull, 1982a). In early broiler growth, heterosis for body weight increases from hatch to slaughter age (Aggarwal et al., 1979: Roy and Kumar, 1981). Heterosis for egg production traits in a 2nd production cycle was greater than or equal to that in the 1st cycle. Reciprocal differences for HHP and egg yield increased in the 2nd cycle, but were unchanged for most other traits (Gowe and Fairfull, 1982a).

Table 6 shows correlations between performance in the 1st and 2nd produc­tion cycles of 24 Leghorn crosses at 1 and 3 hens/cage for VIAB, HHP and EWT. Performance in 1st and 2nd cycles is correlated. However, imperfect correla­tions suggest changes in ranking of the crosses. Taken with the results shown in Table 5, there is a tendency for the better crosses to be better whether

247

kept over a second year or crowded several to a cage as is done in industry.

TABLE 5. Correlation for mean VIAB, HHP and EWT between hens housed 1 and 3/cage for 24 strain crosses

133-496d 546-909d

VIAB 0.18 0.48HHP 0.72 0.54EWTa 0.72 0.63

aAt 240 d and 690 d.

TABLE 6. Correlation between performance in 1st and 2nd cycles of 24 strain crosses housed 1 and 3 hens/cage (h/c)

Trait 1 Trait 2 1 h/c 3 h/c

VIAB 133-496d VIAB 546-909d 0.50 0.51HHP 133-496d HHP 546-909d 0.68 0.68EWT 240d EWT 690d 0.69 0.87

Selection

Cross performance can be improved genetically in several ways. Inbred lines can be developed and tested with better nicking lines being retained. Recipro­cal recurrent selection (RRS) can be used to maximize heterosis. Selection within lines and crossing the improved pure lines can also be a valid proce­dure. As far as the authors know, each of these methods with variations has been used by international breeding companies. However, it is likely that most commercial stocks are improved by selection within strains with very few still using a form of RRS, and if a strain crossing program is used it is used with a strong within strain selection program. Bell (1982) recently reviewed selection for heterosis in laboratory and domestic species including poultry. He also presented the results of a commercial RRS program with egg stocks. Although RRS has been compared with within-strain selection procedures in poultry (Saadeh et_ al., 1968: Calhoon and Bohren, 1974), the results were not conclusive. The critical experiment remains to be done.

Not much has been published on long-term within-line selection and how the heterosis generated in crossing the selected lines changes as selection pro­gresses. The long-term selection studies of Cole and Hutt (1973) and Gowe (Gowe at al., 1973: Gowe and Fairfull, 1985) are germane. Both studiesincluded selection for several economic traits including egg production, via­bility and egg weight as well as other traits. Also in both studies, the selected strains were crossed periodically to assess the merit of the crosses. The results in Table 7 are from Gowe and Fairfull (1982b). Corrected for envi­ronmental trend (using control strain 7), there were genetic gains of 18 eggs and -150 g body weight (BWT), and virtually no change in egg weight (which was the goal at that time). There was substantial heterosis for these traits with no sign of a decrease over generations. As discussed earlier, heterosis would be expected to vary somewhat with environment, but more extensive results gen­erally confirm the result observed (e.g. Fairfull et y.., 1984, 1986). In addition, selection can result in substantial heterosis in very few generations in lines derived from the same strain (Gowe and Fairfull, 1982a: Fairfull et al., 1983). Cole and Hutt (1973) also found significant heterosis in crosses of selected lines that remained constant in magnitude as the lines improved. In their study, the pure strains laid 49 eggs more in 1969 than in 1948-49 (a 31% increase). Crosses were produced in these years and in 1959 with heterosis ranging from 10 to 13%.

In any generation, gains from non-additive effects or crossing (e.g. 20+ eggs) can be expected to be greater than those due to additive effects or

248

selection (e.g. 1-2 eggs). This difference would be smaller for growth. How­ever, gains due to additive effects accumulated over 20-30 generations can be expected to be at least 30-40% of pure strain performance for egg production, a trait with low heritability, in a multi-trait selection program. in meat birds, the additive gain for growth would be greater. Chambers et al. (1981) estimate carcasses of modern broilers are more than twice (130%) the size of those of 20 years ago at present slaughter age. Also, there are few if any indications that additive genetic variation for growth or for egg production is being exhausted. Thus it is hard to believe that breeding systems that give tangibly less emphasis to additive genetic effects would ultimately surpass one that gives heavy emphasis to additive variation. Breeding systems that emphasize additive effects can also utilize non-additive effects in a crossing program. It remains to be demonstrated that selection for crossing ability (RRS) will achieve better results than selecting for within strain performance and crossing the selected strains. Such an experiment must encompass many generations to be conclusive. RRS breeding schemes do not do as well in the earlier generations as within-strain selection programs (Bell, 1982).

TABLE 7. Mean performance in 1972 and 1977, and the difference (D),with heterosis (H%) and reciprocal difference (R%) expressed as apercentage of the mid--parent value (from Gowe and Fairfull, 1982b)

HHP TO 497d EWT at 240d BWT at 365d1972 1977 D 1972 1977 D 1972 1977 D

la 251 258 8 56.6 56.3 -0.3 2.09 1.85 -0.248a 237 246 9 57.9 57.1 -0.8 1.92 1.70 -0.22

1x8 264 292 28 56.9 57.3 0.4 2.09 1.83 -0.268x1 234 269 35 59.5 58.0 -1.5 2.11 1.84 -0.27

7 217 207 -10 54.0 53.7 -0.3 1.83 1.75 -0.08

H% 2% 11% 1.7% 1.7% 4.7% 3.4%R% 12% 9% 4.5% 1.2% 1.0% 0.6%

aln 1972, strains 1 and 8 were in a separate building from the strain crosses and control strain 7.

A more unified approach to the use of additive and non-additive genetic variation is required. New theories need to be proposed and tested. Kinghorn (1982) has proposed one such approach limited to cases where dominance is the important non-additive effect. In view of our relatively poor understanding of multi-trait selection, it is to be hoped that many more attempts will be made.

Specialized Sire and Dam Lines

The exploitation of heterosis is not the only rationale for the use of crosses. Where the functions of dam (reproduction) and progeny (growth and carcass quality) can be separated such as in the production of meat, the efficiency of the system can be maximized by using specialized sire and dam lines (Smith, 1964). This is achieved in part by maximizing selection pressures and genetic gains due to additive effects.

The modern broiler industry is so competitive that a high proportion of the selection pressure has been devoted to growth with less selection for other traits. However, most broiler dam lines are being selected for egg production

249

and egg size, and perhaps, age at sexual maturity and shell quality traits as well. The commercial meat industry is using an approach of specialized sire and dam lines. Unfortunately most of the knowledge gained is not common and because of commercial restraints opportunities for comparisons of methods are restricted. Quality research in this area is needed for all meat birds.

Fatness and feed efficiency are now issues in broiler breeding. Generally, breeders have relied on correlated improvement in feed efficiency due to se­lection for rapid growth. Recent research challenges the efficacy of this approach (see Fairfull and Chambers, 1984). Also, the modern broiler is a very fast growing bird that achieves very large size at an early age. The size of these immature birds puts enormous pressure on ligaments and tendons contributing to an increasing number of skeletal problems (Riddell, 1981). There will be an increasing need to emphasize freedom from crippling in sire lines in particular. Some breeders are dealing with crippling, fatness and feed efficiency in breeding programs for meat birds. However, major questions remain as to approaches.

Prediction of cross performanceExpected cross performance defined in terms of average genetic effects of

lines (g1), individual heterosis due to dominance (h1) and to epistasis (hE), as well as maternal (gM ), maternal sex-linked (gw ) and paternal (gz) effects, are shown below for males and females. See Dickerson (1973) and Fewson (1974) for other definitions. In the models that follow, maternal (hM ) and paternal (hp ) heterosis are not presented. To add them see Dickerson (1973). The expectations of heterosis due to epistasis in two line combinations (i.e. hgB) follow Sheridan (1981). Expectations for three line combinations and higher fi.e. h^(BC)^ are pooled since there is more than one combination that could potentially result in heterosis. These models are are specific to poultry for a single environment. The sire line or cross appears first in any cross designation (i.e. 9 AxB= female progeny of an A line sire bred to a B line dam).

9 AxB- t tg f + Vigl + g? + g " + g " + h j L + hE'ab<f AxB= * g j + ttgl + *gl + %g| + 9B + h*B + h ^

9 a x (b c )= ‘Agl + ‘Agl + ‘Agl + % + g" + g" + + y*hL + v‘hAC + ‘AhE“AC a (b c )’ yA yB yC yA yC yBC AB AC AB AC<J Ax ( bc ) = ‘Agl + ‘Ag* + ‘Ag* + ‘Agl + ttg® + g" + ‘Ah* + ‘Ah* + ‘Ah® + V.hE + ■AhEAC A (BC)B yBC

9 (a b )x c= ‘A g ^ ‘AgB+‘Ag^+‘Ag^‘Ag^g”+g”+‘AhJc+‘AhgC+,Ahftc+‘AhBC+‘,4h (AB)C* ( AB) *C= ,/*gR+V.g^+‘AgJ+‘AgV‘Ag|+‘AgJ+g”+‘AhJc+‘Ah^c+‘AhJc+y.h®c+‘Ah^AB )c

9 ( ab ) x (CD) - %g* + %g* + ‘Ag* + ‘Ag* + ‘Ag^ + ‘AgB + g” + g”D + ‘Ah*c + ‘Ah*D + ‘Ah*c +

1,4hBD + 1/32(hA(CD) + hB(CD) + h(AB)C + h(AB)D) + V*h(AB)(CD) t ( ab ) x ( cd ) — ‘Ag* + %g* + ‘Ag* + ‘Ag* + ‘Ag^ + ‘Ag| + ‘Ag£ + %g^ + g”D + ‘Ah*c + ‘A h ^

+ ‘AhBC + y‘hBD + 1/32(hA(CD) + hB(CD) + h (AB)C + h (AB)D) + ‘̂ (AB) (CD)Paternal effects in poultry are due to Z-chromosome linked genes (gz).

Although a true "paternal" effect (gp ) due to semen transmitted disease or250

other effect is possibile, there is little evidence for it and it is probably negligible. Maternal effects (gM ) are usually confounded with effects due to w-chromosome linked genes (gw ) in females. Where both sexes can be measured, maternal and sex-linked effects can usually be separated unambigu­ously. For the 3-way cross Rx(BC), the Z-chromosomes would be parental types whereas in the (AB)xC cross one generation of recombination would be possible for the sire (AB) Z-linked genes. In specific cases, the model could be sim­plified. For body weight, egg weight and sexual maturity, the hE terms could be deleted from the model. For specific gravity and Haugh units, all heterosis components could be dropped. For female sex-limited traits where gw and gM cannot be separated, they could be defined as gM , etc. The h?AB)C* hI(BC) and higher order terms are likely small but occasion­ally could be highly significant. Coefficients of the terms for heterosis due to epistasis depend on how it is defined. See Dickerson (1973), Sheridan (1980) and Hill (1982) for models and theory. Where relevant data exists, the model should be extended to include an environmental effect (ej) and interaction terms (i.e. '^gjeji) so that the importance of genetic effects across environments could be modelled and examined empirically.

Predictions of cross performance can be lower order cross means (see Hill and Nordskog,

TABLE 8. Correlation of strain cross data with expected values

made from parental strain and 1958). Expected components of the predicted means based on the models (i.e. ‘AA + ‘AB = ‘AgJ + ‘Agl + ‘Agft + 'Aqf + ‘AgJ +

+ ‘AqH +

Fitted PuresExpected from

2-Way 3-Wavmatched with those of the cross predicted (i.e. AxB = ‘AgJ +%gl + + g g + g§ + h£B +

VIABa -0.22 bfe) t° give an idea of the2-Way HHPa 0.39 value and validity of some com-

A50P 0.90 ponents. Table 8 presents cor-BWT 0.86 relations between actual and

expected performance based onVIAB 0.08 0.09 four pure strains and crosses

3-Way HHP 0.21 0.67 from contemporaneous LeghornA50P 0.79 0.90 strain crosses of Fairfull etBWT 0.70 0.81 al. (1986). Expectations were

derived as follows: from pureVIAB - 0.01 -0.43 strains, E(AxB) = ‘AA+ 'AB,

4-Way HHP - 0.30 0.27 E(Ax{BC}) = ‘AA + ‘AB + *AC,A50P - 0.53 0.52 E({AB}x{AB}) = ‘AA + ‘AB: fromBWT - -0.04 0.43 2-way crosses, E(Ax{BC}) = ‘AAxB

+ *AAxC, E({AB}x{CD}) = ‘AAxC +VIAB 0.31 -0.03 0.37b ‘AAXD + ‘ABXC + ‘ABxD, E({AB} X

f2 HHP 0.53 0.69 0.75b {AB}) = ‘AAxB + ‘ABxA; fromA50P 0.92 0.94 0.94b 3-way crosses, E({AB} x {CD}) =BWT 0.83 0.89 0.91b ‘A Ax(CD) + ‘ABx(CD): from pure

strains and 2-way crosses.aFrom 134 to 497 d. E ({AB} X {AB}) =~*AA + ‘AB +“Pure strains and 2-way used. ‘AAxB. Expectations based on

pure strain performance pre­dicted 2-way and ?2 cross means of sexual maturity (A50P) and body weight (BWT) well, and means of 3-ways tolerably well showing the importance of additive genetic effects for these traits. The very poor prediction of BWT in the 4-way crosses from 2-way means would not be expected. Matched 3-way crosses improved the prediction of BWT in 4-way crosses but did not otherwise

251

improve upon prediction with 2-way crosses. For HHP, pure strain performance does not seem to be a good predictor of cross performance probably due to the importance of espistasis for HHP. Matched 2-way crosses generally improved the prediction of HHP although HHP was not predicted as well as A50P and BWT. Based on our model, 3-way crosses would be expected to predict HHP in 4-way crosses as well as or better than expectations based on 2-way crosses, but no improvement was realized. LHM was poorly predicted. In general 4-way crosses were not predicted as well as 3-way and F2 crosses and this would be expected under our model especially for traits where epistasis was important. Eisen et al. (1967a,b) predicted topcross and double cross performance. Egg and body weight were predicted best. At present, the pre- diction of cross performance is of limited benefit in practical applications in poultry: however, theory is important. Present models are more accurate for traits such as body weight than for egg production and need to be improved.

PRESERVATION OF STRAINS OR BREEDS

This important topic deserves more consideration than can be given here, but is too important not to mention in passing. Strains or breeds that have little or no present economic value are still useful (Maijala, 1974: Makarechian, 1974: Shoffner, 1974: Crawford, 1984). The total gene pool of commercial stocks is relatively narrow as similar types of birds are used commercially and genes are interchanged among breeders. The preservation of non-economic strains or breeds serves as a form of insurance against economic change or for times of crisis such as climatic events, war or pestilence. Included in these stocks are mutant, marker gene and chromosome aberration lines. Such lines are the raw material of basic and applied research. Application of this knowledge is often wider than poultry science as discov­eries are often applicable to other species. Now that the age of gene trans­fer is upon us, oddities, collected mutant lines and others will serve not only as a source of genes for transfer, but also as a source of inspiration with nature suggesting that which man might never think of. Finally, many local strains or breeds could serve as sources to combine local adaptation with high production potential.

REFERENCES

AGGARWAL, C.K., MOHAPATRA, S.C., SINHA, S.P. SHARMA, P.N., and AHUJA, S.D. 1979. Estimation of combining ability in broilers from a full diallel cross. Br. Poult, sci. 20, 185-190.

AL-MURRANI, W.K. 1978. Maternal effects on embryonic and post-embryonic growth in poultry. Br. Poult. Sci. 19, 277-281.

BACON, L.D., SMITH, E.J., and CRITTENDEN, L.B. 1985. Association of genes determining slow feathering (K) and an endogenous virus (ev 21) on the 7,- chromosome. Poult. Sci. 64(Suppl. 1), 60(Abstr).

BARLOW, R. 1981. Experimental evidence for interaction between heterosis and environment in animals. Anim. Breed. Abstr. 49, 715-737.

BELL, A.E. 1982. Selection for heterosis - results with laboratory anddomestic animals. In: 2nd Wrld Congr. Genet. Appl. Livest. Prod. 6, 206-227.

BERNON, D.E., and CHAMBERS, J.R. 1985. Maternal and sex-linked genetic effects in broiler parent stocks. Poult. Sci. 64, 29-38.

252

BLYTH, J.S.S., and SANG, J.H. I960. Survey of line crosses in a Brown Leghorn flock. 1. Egg production. Genet. Res. 1, 408-421.

CALHOON, R.E., and BOHREN, B.B. 1974. Genetic gains from reciprocal recurrent and within-line selection for egg production in the fowl. Theor. Appl, Genet. 44, 364-372.

CHAMBERS, J.R., GAVORA, J.S., and FORTIN, A. 1981. Genetic changes in meat-type chickens in the last twenty years. Can. J. Anim. Scl. 61, 555-563.

COLE, R.K., and HUTT, F.B. 1973. Selection and heterosis in Cornell White Leghorns: A review, with special consideration of interstrain hybrids. Anim. Br. Abstr. 41, 103-118.

CRAWFORD, R.D. 1984. Assessment and conservation of animal genetic resources in Canada. Can. J. Anim. Sci. 64, 235-251.

CRITTENDON, L.B., SMITH, E.J., and FADLY, A.M. 1984. Influence of endogenous viral (ev) gene expression and strain of exogenous avian leukosis virus (ALV) on mortality and ALV infection and shedding in chickens. Avian Pis. 28,1037-1056.

DICKERSON, G.E. 1965. Experimental evaluation of selection theory in poultry. Genet■ Today 3, 747-761.

DICKERSON, G.E. 1973. Inbreeding and heterosis in animals, pp. 54-77. In: Proc. Anim, Breed. Genet. Svmp. Honor J.L. Lush, Amer. Soc. Anim. Sci.,Blacksburg.

EISEN, E.J., BOHREN, B.B., McKEAN, H.E., and KING, S.C. 1967a. Prediction of topcross performance from inbred performance in poultry. Poult. Sci. 46,195-203.

EISEN, E.J., BOHREN, B.B., McKEAN, H.E., and KING, S.C. 1967b. Combiningability among single crosses and predicting double cross performances in poultry. Br. Poult, sci. 8, 231-242.

EMSLEY, A., DICKERSON, G.E., and GOWE, R.S. 1980. Inbreeding effects onlayer performance at two levels of protein intake. Poult. Sci. 59, 1155-1166.

FADLY, A.M., and OKAZAKI, W. 1982. Studies of avian leukosis virus infection in chickens from a commercial breeder flock. Poult. Sci. 61, 1055-1060.

FAIRFULL, R.W., and CHAMBERS, J.R. 1984. Breeding for feed efficiency: Poultry. Can. J. Anim. Sci. 64, 513-527.

FAIRFULL, R.W., GOWE, R.S., and EMSLEY, J.A.B. 1983. Diallel cross of Six long-term selected Leghorn strains with emphasis on heterosis and reciprocal effects. Br. Poult. Sci. 24, 133-158.

FAIRFULL, R.W., GOWE, R.S., and NAGAI, J. 1986. Dominance and epistasis in heterosis of White Leghorn strain crosses. Can. J. Anim. Sci. (submitted).

FAIRFULL, R.W., SPENCER, J.L., GAVORA, J.S., and GOWE, R.S. 1984. Performance of laying hens positive for group specific viral antigen or for

253

lymphoid leukosis virus in egg albumen and egg yolk. Archiv fur Geflugelk. 48, 187-192.

FEWSON, D. 1974. Crossbreeding in farm animals, pp. 25-38. In: Proc. Work. Symp. Breed. Eval. Cross. Exp'ts. Farm Anim., Res. Inst. Anim. Husb., Zeist.

FLOCK, D.K. 1980. Heterosisschatzungen in einer population von weissenLeghorn nach langjahriger RES. In: 6th Europ. Poult. Conf. 2, 57-63.

FRIARS, G.W., JEROME, F.N. , WEEDEN, L.T., and ASHTON, G.C. 1963. Body weights and reproduction traits of two strains and reciprocal crosses of broad breasted Bronze Turkeys. Poult. Sci. 42, 935-940.

GAVORA, J.S., and SPENCER, J.L. 1985. Effects of lymphoid leukosis virus infection on response to selection, pp. 17-24. In: Poult. Genet. Breed., Ed. Hill, W.G., Manson, J.M., and Hewitt, D. Br. Poultry Sci., Edinburgh.

GOWE, R.S., and FAIRFULL, R.W. 1982a. Heterosis in egg-type chickens. In: 2nd Wrld Congr. Genet. Appl. Livest. Prod. 6, 228-242.

GOWE, R.S., and FAIRFULL, R.W. 1982b. Some lessons from selection studies in poultry. In: Proc. Wrld Congr. Sheep and Beef Cattle Breed. 1, 261-281.

GOWE, R.S., and FAIRFULL, R.W. 1985. The direct response to long-termselection for multiple traits in egg stocks and changes in genetic parameters with selection. pp. 125-146. In: Poult. Genet. Breed., Ed. Hill, W.G.,Manson, J.M., and Hewitt, D. Br. Poultry Sci., Edinburgh.

GOWE, R.S., LENTZ, W.E., and STRAIN, J.H. 1973. Long-term selection for egg production in several strains of White Leghorns: Performance of selected and control strains including genetic parameters of two control strains, pp. 225-245. In: 4th Europ. Poult. Sci. Conf., London.

HALE, R.W., and CLAYTON, G.A. 1965. A diallel crossing experiment with two breeds of laying fowl. Br. Poult. Sci. 6, 153-174.

HARRIS, D.L., GARWOOD, V.A., LOWE, P.C., HESTER, P.Y., CRITTENDEN, L.B., and FADLY, A.M. 1984. Influence of sex-linked feathering phenotypes of parents and progeny upon lymphoid leukosis virus infection status and egg production. Poult. Sci. 63, 401-413.

HILL, J.F., and NORDSKOG, A.W. 1958. Heterosis in poultry. 3. Predicting combining ability of performance in the crossbred fowl. Poult. Sci. 37,1159-1169.

HILL, W.G. 1982. Dominance and epistasis as components of heterosis. z. Tierzuchtg. Zuchtgsbiol. 99, 161-168.HORN, P. 1974. Comparative evaluation of the hatchability of single- and three-way cross combinations in meat type fowl. Acta Agron. Sci. Hung. 23, 45-55.

HORN, P. 1982. Performance test of progeny groups of paternal half sib pure and crossbred laying hens in optimal and suboptimal environments in two cycles of production. Allatteny. Takarm. 31, 555-566. (English summary)

254

HORN, P., TRINH, D.D., and KALLAY, B. 1980. Heterosis in optimal and suboptimal environment in layers during the first and second laying period after force moult. In: 6th Europ. Poult. Conf. 2, 48-56.

HULL, P., GOWE, R.S., SLEN, S.B., and CRAWFORD, R.D. 1963. A comparison of the interaction, with two types of environment, of pure strains or strain crosses of poultry. Genet. Res. 4, 370-381.

JEFFERS, T.K., CHALLEY, J.R., and McGIBBON, W.H. 1970. Response of several lines of fowl and their single-cross progeny to experimental infection with Eimeria tenella. Avian Pis. 14. 203-210.

JEROME, F.N., FRIARS, G.W., and WEEDEN, L.T. 1960. Heterosis for growth in turkeys. Poult. Sci. 39, 1225-1226.

KINGHORN, B. 1982. Genetic effects in crossbreeding. II. Multibreed selection indices. Z. Tierzuchtg. Zuchtgsbiol. 99, 315-320.

KROSIGK, C.M. von, McCLARY, C.F., VIELITZ, E. , and ZANDER, D.V. 1972.Selection for resistance to Marek's disease and its expected effects on other important traits in White Leghorn strain crosses. Avian Pis. 16, 11-19.

LOWE, P.C., and GARWOOD, V.A. 1981. Independent effects of K and k+ alleles and maternal origin on mortality and performance of crossbred chickens. Poult. Sci. 60, 1123-1126.

MAIJALA, K. 1974. Conservation of animal breeds in general. In: 1st Wrld Congr. Genet. Appl. Livest. Prod. 2, 37-46.

MAKARECHIAN, M. 1974. Evaluation of the genetic potentialities of native breeds of poultry for egg production. In: 1st Wrld Congr. Genet. Appl. Livest. Prod. 3, 1207-1210.

MERRITT, E.S., and GOWE, R.S. 1960. Combining ability among breeds and strains of meat type fowl. Can. J. Genet. Cvtol. 2, 286-294.

MERRITT, E.S., and GOWE, R.S. 1965. Post embryonic growth in relation to egg weight. Poult. Sci. 44, 477-480.

MORRIS, J.A., and BINET, F.E. 1966. Heterosis and selection in the domestic fowl. Aust. J. Biol. Sci. 19. 575-585.

NESTOR, K.E. 1971. Genetics of growth and reproduction in the turkey. 4. Strain crossing for improvement of growth and reproduction. Poult Sci 50 1683-1689. 1

NORDSKOG, A.W., and PEVZNER, I.Y. 1977. Sex-linkage versus maternal anti­bodies in the genetic control of disease. Wrld's Poult. Sci, j. 33. 21-30.

PAYNE, L.N., HOLMES, A.E., HOWES, K., PATTISON, M., POLLOCK, D.L., and WALTERS, D.E. 1982. Further studies on the eradication and epizootiology of lymphoid leukosis virus infection in a commercial strain of chickens. Avian Pathol. 11, 145-162.

PIRCHNER, F. 1969. Population Genetics in Animal Breeding. Freeman and Co., San Francisco.

255

PODGER, R.N., YOO, B.H.t and SHELDON, B.L. 1980- Cross—breeding performance of White Leghorn and Australorp lines selected for short interval between eggs under continuous light. 3. Effects of production and maintenance variables on feed intake. pp. 71-76. In: Proc. S. Pacif. Poult. Sci. Conv.. Wrld's Poult. Sci. Assoc., Auckland.

RIDDELL, C. 1981. Skeletal deformities in poultry. Adv. Vet. Sci. Comp. Med. 25, 277-310.

ROY, T.C., and KUMAR, J. 1981. A study on the pattern of genetic variation in the progenies of a 3x3 diallel cross of commercial broiler strains of chicken. Ind. Vet. J. 58, 461-464.

SAADEH, H.K., CRAIG, J.V., SMITH, L.T., and WEARDEN, S. 1968. Effectiveness of alternative breeding systems for increasing rate of egg production in chickens. Poult. Sci. 47, 1057-1072.

SHERIDAN, A.K. 1980. A new explanation for egg production heterosis in crosses between White Leghorns and Australorps. Br. Poult. Sci. 21. 85-88.

SHERIDAN, A.K. 1981. Crossbreeding and heterosis. Anim. Breed. Abstr 49. 131-144.

SHERIDAN, A.K., and RANDALL, M.C. 1977. Heterosis for egg production in White Leghorn Australorp crosses. Br. Poult. Sci. 18. 69-77

SHOFFNER, R.N. 1974. Usefulness of chromosomal abberations in animal breeding. In: 1st Wrld Congr. Genet. Appl. Livest. Prod. 1, 135-149.

SMITH, C. 1964. The use of specialized sire and dam lines in selection formeat production. Anim. Prod. 6, 337-344.

TARAN, M., GABRIEL, E., MOSES, E., and SOLLER, M. 1972. Performance of purebred and crossbred progeny of White Leghorn and New Hampshire sires. Br. Poult. Sci. 13. 331-339.

TIJEN, W.F. van 1977. Shell quality in poultry as seen from the breeder's viewpoint. 2. Search for heterosis effects. Poult. Sci. 56, 1115-1120.

WEARDEN. S., TINDELL, D., and CRAIG, J.V. 1965. Use of a full diallel cross to estimate general and specific combining ability in chickens. Poult. Sci. 44, 1043-1053.

WILLEKE, H. 1982. Heterosis of different egg-laying traits, in reciprocal crosses of brown egg-layers. In: 2nd Wrld Congr. Genet. Appl. Livest. Prod. 6j_ 841-846.

YOO, B.H., SHELDON, B.L., and PODGER, R.N. 1984. Cross—breeding performance of White Leghorn and Australorp lines selected under continuous light for short interval between eggs: Full year egg production and efficiency of food utilisation. Br. Poult. Sci. 25, 233-243.

256