Inbreeding is the mating of individuals who are related by descent, leading to an increase in the frequency of homozygotes and decrease of heterozygotes. The negative effect of inbreeding, known as inbreeding depression is a reduction in the mean of quantitative traits which is either related to reproduction capacity or physiological efficiency and is a consequence of recessive deleterious alleles (Miranda Filho 1999). Homozygosity permits the expres-sion of recessive alleles that may have been masked by a dominant allele under heterozygosity. The exposure of the recessive alleles often results in a reduction in individual performances, which is referred to as inbreeding depression or loss of vigour. Inbreeding depression of some populations may be so severe as to prevent their usefulness (Ajala 1992). The potential of some populations as sources of high- yielding inbred lines is largely dependent on the inbreeding depression observed for quantitative traits. Of the different systems of inbreeding (i.e. Selfing, Half-sib, Full-sib and Backcrossing), homozygosity is attained faster through selfing because 50% homozygosity can be attained in one generation compared to three generation for full-sib to attain the same level of homozygosity. The estimation of inbreeding depression is obtained faster through the com-parison of the population per se (S₀) and their selfed (S₁) counterparts. The rate of attaining homozygosity has been shown to also vary with different traits. Fakorede et al. (1993) compared the rate of inbreeding in five (S₁-S₅) generations of inbreeding in TZSR-W maize population with the parental variety (S₀) for yield and agronomic traits. Means obtained for each generation showed differential effects of inbreeding on the traits considered, with emer-gence percentage (E%), emergence index (EI), emergence rate index (ERI), days to 50% tasseling, anthesis and silk-ing respectively recording positive inbreeding depression (+0.2627, +0.0118, +0.1250, +0.0167, +0.0112, +0.0223), which indicate that inbred lines tend to emerge slowly and flower later than the source population. Conversely, the inbreeding (%) at the S₁ and S₄ generation for grain yield, ear number, ear length, ear diameter, stand count, plant height and days to silk were ‒42.3, ‒9.9, ‒12.9, ‒11.6, ‒7.2, ‒14.6 and +2.5 respectively, indicating that grain yield and yield components were the most sensitive to inbreeding. The authors hypothesized that the differences may be due to the level of complexity in the inheritance of different traits.

Meghji et al. (1984), conducted a study, to investigate the levels of variation in inbreeding depression of different maize populations representing three breeding eras (1930s, 1950s and 1970s) which were also evaluated under different plant densities at two locations (Urbana and Shabbona). Except for days to silking, differences in inbreeding depression were observed for all the other traits considered across era. However, greater level of inbreeding depression were recorded for grain yield of maize hybrids from the 1970s, implying that modern maize varieties are more heterozygous at loci that affect the expression of maize grain yield and that improvement of the inbred parents of the hybrids have occurred at different loci. Inbreeding depression for grain yield was higher at high plant density than at low density in the modern maize variety which the authors attributed to the fact that a portion of the inbreeding did occur at loci that conditioned the responsiveness of these varieties to high plant densities. Furthermore, lower inbreeding depression (%) was obtained for plant and ear heights in the hybrids of 1970s compared to those of 1930s and 1950s. The authors therefore at-tributed this to improvement for these traits in the inbred parents at those loci.

Lima et al. (1984) investigated inbreeding depression by evaluating inbred lines generated by selfing from 32 Brazilian maize populations which consisted of OPVs, synthetic and composite varieties in four different locations. Inbreeding depression (%) was higher for grain yield than for other traits and ranged from 27.0 to 57.9% among the different populations. However, inbreeding depression was generally smaller for populations derived from inbred lines than for those derived from OPVs or composite varieties. The authors concluded that the estimates of A and d as a measure of the contribution of homozygotes and heterozygotes, respectively, to the mean performance of populations and the most positive values of A/d are indicative of the potential of the genotypes as sources of high-yielding lines. The authors further con-cluded that the low values of A/d resulted from negative contribution of homozygotes in the population, which also revealed the presence of epistasis, and occurrence of del-eterious recessives in the populations. The objective of this study therefore was to investigate inbreeding depression of some tropical maize varieties developed at different breed-ing eras and evaluated in optimal and stem borer infested condition.

Generation of inbred lines

S₁ progenies were generated from 10 maize population developed between 1979 and 2008 by selfing 35-50 plants in each population during the 2012 cropping season at the International Institute of Tropical Agriculture (IITA), Ibadan, Nigeria. The resultant S₁ progenies were harvested, processed, fumigated and stored in the cold room prior to field evaluation. The characteristic of the tested varieties are found in Table 1.

Table 1 . The characteristics of the ten maize varieties tested.

S/N	Genotype	Year of release	Grain colour	Endosperm type	Maturity rating	Breeding emphasis
1.	TZSR-W-1	1979	White	Flint	Late	Tropically adapted and streak virus resistance
2.	DMR-LSR-W	1980	White	Dent	Late	Downy mildew, low-N and streak virus resistance/tolerance
3.	DMR-LSR-Y	1980	Yellow	Dent	Late	Downy mildew, low-N and streak virus resistance/tolerance
4.	TZSR-Y-1	1979	Yellow	Flint	Late	Tropically adapted and streak virus resistance
5.	ACR99TZLCOMP4DMRSR	1999	White	Dent/Flint	Late	Tropical adapted, low-N and downy mildew tolerance
6.	BR9922DMRSR	2008	White	Flint	Late	Borers, downy mildew and streak virus resistance
7.	BR9928DMRSR	2008	Yellow	Flint	Late	Borers, downy mildew and streak virus resistance
8.	BR9943DMRSR	2008	White	Flint	Late	Borers, downy mildew and streak resistance
9.	AMATZBR-WC2B	2008	White	Flint	Late	Tropically adapted and borers resistance
10.	TZBRELD.4C0W	2000	White	Flint	Late	Tropically adapted and borer resistance

Evaluation of parents and inbred at seven environments

The ten parents and their respective S₁ inbred were evaluated in seven (7) environments. The first environ-ments was under artificial infestation with stem borer at the IITA, Ibadan (Latitude7⁰ 22¹N, Longitude 3⁰ 58¹E), while other environments viz: Ikenne (Latitude 6⁰ 53¹N, Longi-tude 3⁰ 42¹E) and Ile-Ife (Latitude 7⁰ 18¹N, Longitude 4⁰ 33¹E) (rain forest region) were regarded as stress-free environments. Mokwa (Latitude 9⁰ 18¹N, Longitude 5⁰N 4¹E) and Zaria (Latitude 12⁰ 00¹N, Longitude 8⁰ 22¹E) (Guinea savanna) where the genetic materials were eval-uated under high and low N conditions respectively. In all evaluations, two row plots were used. Each row was 6m in length, spaced at 0.75 m between rows 0.25 m within rows with four replications to give a population density of ap-proximately 53,333 plants per hectare. Observed cultural practices included pre-emergence spray of gramozone and primextra for weed control supplemented with hand weeding as necessary during the season. Fertilizer was also split applied using N-P-K 15:15:15 at 10 days after planting (DAP) at the rate of 30 kg N/ha and top dressed with urea six WAP at the same rate. However, the four locations at Mokwa and Zaria were the low and high-N environment with two different levels of nitrogen application (30 kgha^‒1 and 90 kgha^‒1) respectively.

Artificial infestation on Ibadan trial with stem borers

Trials at Ibadan was artificially infested with egg masses of Sesamia calamists at three weeks after emergence. It was again infested with Sesamia at flowering stage with egg masses of Eldana Sachar.

Inbreeding effect measured in percentage (I%) of selfed populations (S₁) response compared to response per se of parent (S₀) were calculated as:

I = 100[(S_O ‒ S₁) / S_O]

Where S_O and S₁ represent the genotype per se and its selfed counterpart, respectively. Means due to S_O and S₁ were expressed as:

S_O = A + d and

S₁ = A + ½ d

Where A represents the potential contribution of homo-zygotes and d is deviation due to the heterozygotes in a population. For each generation of inbreeding, the proportion of heterozygotes in the population will be halved. The values of A and d are dependent on gene frequencies for the population.

I = 100 [(S₀ ‒ S₁) / S₀

d = S₀ ‒ S₁

A = [2(S₁) ‒ S₀] / 2

The most positive values of A/d are a measure of the potential of the populations as sources of high-yielding inbred lines (Lima et al. 1984).

Grain yield for population per se (S₀) varied from 1.2 t/h^‒1 for DMR-LSR-Y to 1.9 t/h^‒1 for BR9922DMRSR, while the S₁ lines varied from 0.92 t/h^‒1 for BR9922DMRSR to 1.49 t/h^‒1 for TZSR-Y-1 and TZBRELD.4C₀W in optimal environment of Ife and Ikenne (Table 2). High inbreeding depression across natural field environments was observed in BR9922DMRSR, AMATZBR-WC2B, ACR99TZLCOMP- 4DMRSR, TZBRELD.4C₀W and BR9928DMRSR, and low in TZSR-Y-1 and TZSR-W-1. The high positive values of A/d that indicates maize populations as sources of genes for the extraction of high-yielding inbred lines were observed in vars. DMR-LSR-Y and TZSR-W-1, while other populations had very low values.

Table 2 . Comparative response of maize inbreeding depression to optimal field condition (Ikenne and Ife).

Population	Era	S0	S1	I	A	D	A/d
TZSR-W-1	1	1.34	1.22	8.96	0.55	0.12	4.58
DMR-LSR-W	1	1.57	1.28	18.47	0.50	0.29	1.71
DMR-LSR-Y	1	1.20	1.19	0.83	0.59	0.01	59.00
TZSR-Y-1	1	1.38	1.49	‒7.97	0.80	‒0.11	‒7.27
ACR99TZLCOMP4-DMRSR	1	2.00	1.37	31.50	0.37	0.63	0.59
BR9922DMRSR	2	2.07	1.28	38.16	0.25	0.79	0.31
BR9928DMRSR	2	1.27	0.92	27.56	0.29	0.35	0.81
BR9943DMRSR	2	1.55	1.27	18.06	0.50	0.28	1.77
AMATZBR-WC2B	2	1.91	1.21	36.65	0.26	0.70	0.36
TZBRELD.4C0W	2	1.87	1.36	27.00	0.43	0.51	0.83
Mean		1.62	1.26	22.09	0.45	0.36	1.26

Values of S₀ represent the genotype per se, S₁ inbreeding effect (I). Where A represents the potential contribution of homozygotes and d is deviation due to the heterozygotes.

However, in low-N environment of Mokwa and Zaria, grain yield in respect of the population per se varied from 1.3 t/h^‒1 for BR9928DMRSR to 1.9 t/h^‒1 for AMTZBR- WC2B while the S₁ lines varied from 0.91 t/h^‒1 for BR9922DMRSR to 1.36 t/h-1 for DMR-LSR-Y and TZBRELD.4C₀W (Table 3). The Inbreeding depression (I) for grain yield ranged from 13% for var. ACR99TZL COMP4-DMRSR to 52% for var. BR9928DMRSR under low-N environment. However, the effect of inbreeding on grain yield was more severe for BR9922DMRSR under low-N condition and least for three populations-BR9943 DMRSR, DMRLSR-Y and ACR99TZLCOMP4DMRSR. Populations-DMR-LSR-Y, ACR99TZLCOMP4DMRSR and BR9943DMRSR had the most positive values of A/d under low-N environment Under high-N environment, grain yield recorded for parental lines varied from 1.23 t/h^‒1 for BR9928DMRSR to 2.15 t/h^‒1 for AMATZBR-WC₂B, while those of the S₁ lines varied from 1.14 t/h^‒1 for DMR-LSR-W to 1.48 t/h^‒1 for DMR-LSR-Y (Table 4). High values of inbreeding depression were observed in AMATZBR-WC₂B, TZBRELD.4C₀W, ACR99TZLCOMP4- DMRSR and BR9922DMRSR and low for TZSR-Y-1, DMR-LSR-Y and BR9928DMRSR across high-N environ-ments. AMATZBR-WC₂B appeared to be the most sen-sitive to inbreeding across high-N environments by record-ing 35.35. The most positive values of A/d were recorded in TZSR-W-1, DMR-LSR-W, DMLSR-Y and TZSR-Y-1 while it was low for ACR99TZLCOMP-4DMRSR, BR9922D MRSR, AMATZBR-WC₂B and TZBRELD.4C₀W across high-N environment.

Table 3 . Comparative response of maize inbreeding depression to optimal field condition Low-N environment (Mokwa & Zaria).

Population	Era	S0	S1	I	A	d	A/d
TZSR-W-1	1	1.73	1.22	29.47	0.35	0.51	0.69
DMR-LSR-W	1	1.65	1.15	30.30	0.32	0.50	0.65
DMR-LSR-Y	1	1.60	1.36	15.00	0.56	0.24	2.33
TZSR-Y-1	1	1.79	1.32	26.25	0.42	0.47	0.90
ACR99TZLCOMP4-DMRSR	1	1.50	1.3	13.33	0.55	0.2	2.75
BR9922DMRSR	2	1.90	0.91	52.10	‒0.04	0.99	‒0.04
R9928DMRSR	2	1.29	0.94	27.13	0.29	0.35	0.84
BR9943DMRSR	2	1.37	1.19	13.13	0.50	0.18	2.80
AMATZBR-WC2B	2	1.90	1.35	28.94	0.40	0.55	0.72
TZBRELD.4C0W	2	1.70	1.36	20.00	0.51	0.34	1.50
Mean		1.64	1.21	26.35	0.38	0.433	0.89

Values of S₀ represent the genotype per se, S₁ inbreeding effect (I). Where A represents the potential contribution of homozygotes and d is deviation due to the heterozygotes.

Table 4 . Comparative response of maize inbreeding depression to High-N environment (Mokwa and Zaria).

Population	Era	S0	S1	I	A	D	A/d
TZSR-W-1	1	1.60	1.23	23.13	0.43	0.37	1.16
DMR-LSR-W	1	1.50	1.14	24.00	0.39	0.36	1.08
DMR-LSR-W	1	1.55	1.48	4.52	0.71	0.07	10.07
TZSR-Y-1	1	1.34	1.30	2.99	0.63	0.04	15.75
ACR99TZLCOMP4DMRSR	1	1.88	1.27	32.45	0.33	0.61	0.54
BR9922DMRSR	2	1.93	1.32	31.61	0.36	0.61	0.58
BR9928DMRSR	2	1.23	1.17	4.88	0.56	0.06	9.25
BR9943DMRSR	2	1.71	1.42	16.96	0.57	0.29	1.95
AMATZBR-WC2B	2	2.15	1.39	35.35	0.32	0.76	0.41
TZBRELD.4C0W	2	2.09	1.36	34.93	0.32	0.73	0.43
Mean		1.70	1.31	21.08	0.46	0.39	1.18

Values of S₀ represent the genotype per se, S₁ inbreeding effect (I). Where A represents the potential contribution of homozygotes and d is deviation due to the heterozygotes.

Similar to the results obtained in other environments, grain yield for parental (S₀) population were higher rela-tive to those of the S₁ lines except for var. BR9928DMRSR under borer infested condition (Table 5). Inbreeding depres-sion under borer infested condition ranged from 1.6% for var. BR9943DMRSR to 68% for var. DMRLSR-Y. Most positive valueof A/d were observed in vars. BR9943DMRSR, TZSR-W and BR9928DMRSR under borer infestation condition.

Table 5 . Comparative response of maize inbreeding depression to stem borer artificial infestation environment (Ibadan).

Population	Era	S0	S1	I	A	D	A/d
TZSR-W-1	1	2.03	1.91	5.78	0.90	0.12	7.65
DMR-LSR-W	1	2.99	2.05	31.38	0.56	0.94	0.59
DMR-LSR-W	1	1.63	0.52	68.26	‒0.30	1.11	‒0.27
TZSR-Y-1	1	2.93	1.06	63.95	‒0.41	1.87	‒0.22
ACR99TZLCOMP4-DMRSR	1	2.44	0.97	60.18	‒0.25	1.47	‒0.17
BR9922DMRSR	2	2.81	1.39	50.56	‒0.02	1.42	‒0.01
BR9928DMRSR	2	2.21	2.56	‒15.95	1.46	0.35	‒4.13
BR9943DMRSR	2	2.48	2.44	1.63	1.20	0.04	29.59
AMATZBR-WC2B	2	2.74	1.11	59.52	‒0.26	1.63	‒0.16
TZBRELD.4C0-W	2	2.38	1.59	33.15	0.40	0.79	0.51
Mean		2.46	1.56	35.85	0.33	0.33	0.90

Values S₀ represents the genotypes per se, S₁ is the inbreeding effect (I) Where A represents the potential contribution of homozygotes and d deviation due to the heterozygotes.

Inbreeding in maize causes reduction in vigor, structure and productivity. Report from Hallauer and Miranda Filho (1981) indicated that the decrease in important traits, such as grain yield resulting from inbreeding can be so severe as to limit the usefulness of such line(s) in a hybrid pro-gramme.

All the maize populations used in this study have been under selection. The objective of the selection programmes has been to increase the frequency of the more favorable alleles for the traits under selection. Across all the en-vironments used for this study, estimates of inbreeding depression (I) for grain yield which ranged from 15.63% for optimum environment to 35.85 under stem borer infestation, showed differences in the severity of the effects of practicing inbreeding in each of the populations and the different environments. For example, the highest values of inbreeding depression for grain yield were recorded among the maize varieties under stem borer infestation. Theoret-ically, expected inbreeding depression for grain yield after one generation of selfing is 50% which was similar to values obtained from this study in BR9922DMRSR with 52% and 50% inbreeding percentage under low-N and stem borer infestation respectively after one generation of selfing.

Except in the borer infested environment, increase in S₀, S₁ and inbreeding depression were observed for the genotypes in more recent era, which is similar to reports of Lamkey and Smith (1987) who investigated the performance and inbreeding depression of populations representing seven eras of maize breeding. The authors also observed increased in inbreeding rate among the more recent eras of maize varieties, coupled with the increase performance of the S₀ and S₁ populations. Consequently, the authors hypo-thesized that the favourable allele frequencies were below 0.50 which must have been increasing progressively with inbreeding or that the more recent eras of maize popu-lations were segregating at more loci. In a similar view, Meghji et al. (1984) reported that modern maize varieties are more heterozygous at loci that affect the expression of maize grain yield than for the maize varieties of previous eras and that the improvement of the parent varieties must have occurred at different loci. One of the major objectives of the selection programmes has been to increase the frequency of the favorable alleles for the traits under selection. The report obtained from this study therefore indicates that maize breeding efforts in the tropic have been effective at increasing the frequency of favorable alleles for grain yield.

The effect of inbreeding was the most severe for var. DMR-LSR-W under borer infestation and least for DMR- LSR-Y in stress free environment. This is an indication that the responses of the maize populations to inbreeding as well as the rate of attaining homozygosity differed with environments. Fakorede et al. (1984) also observed that grain yield and yield components were the most sensitive to inbreeding which according to the authors may be due to be due to the level of complexity in the inheritance of these traits.

The highest inbreeding depression for grain yield was also observed among recently released maize varieties (Era 2) under borer infestation. Meghji et al. (1984) reported similar greater inbreeding depression for grain yield at high plant density than at low density in the modern maize varieties which the authors attributed to a portion of the inbreeding having occurred at loci that condition the responsiveness of the maize varieties to stress condition. Across the four different environments under which the genotypes were evaluated, average inbreeding depression for grain yield were greater relative to other traits con-sidered which should be expected since grain yield is a quantitatively inherited trait, governed by many genes each with minor effects.

According to Lima et al. (1984) the most positive A/d indicates the potential of the genotypes as sources of high- yielding inbred line. The most positive ratio of A to d were recorded by parental varieties DMR-LSR-Y, ACR99TZL COMP4-DMRSR and BR9943DMRSR (Low-N) and DMR- LSR-Y, TZSR-Y-I, BR9928DMRSR (High N), DMR- LSR-Y, TZSR-Y-1, BR9928DMRSR), (TZSR-W-I, DMR- LSR-Y, BR9943DMRSR (stress free) and TZSR-W-I, DMR-LSR-W, Therefore, these parental lines are promising candidates for the extraction of high-yielding inbred lines for formation of hybrids for the specific ecologies and/or environments. The goal of selection programmes has been to increase the frequency of the more favourable alleles for the traits under selection because if selection is effective, there will be increase in the frequency of the desirable alleles, and reduction in inbreeding depression (Scheffler et al. 2008). Most of the old varieties were observed to respond less to inbreeding across the four environments which suggests that most of the loci were probably fixed due to selection.