For agronomic evaluation in respect of vegetative vigour, data was collected from 20 plants for each progeny (five plants per replication) for three crop seasons till the bush canopy was totally covered. For assessment of vegetative vigour, data on the growth parameters such as stem girth, bush spread, number of primary branches, length of the longest primary, number of nodes per primary, inter nodal length and yield component characters such as number of bearing nodes per primary, number of fruits per node was recorded. The fruit yield per plant and progeny yield was recorded during harvest season i.e., Nov. - Dec. for four crop seasons from 2017 to 2020. Data was also collected from Chandragiri and Sln.10 the parents.

Table 2
Weather data pertaining to the study period 20172020

Tem Average temperature, Rainfall - rainy days in parenthesis, RH-Relative Humidity

Analysis of variance was carried out for each character as suggested by Gomez and Gomez (1984). Wherever the treatment differences were found significant, Critical Differences (CD) were worked out at five per cent probability level and values were furnished. The treatment differences that were not significant are indicated as ‘NS’.

Relative heterosis and heterobeltiosis manifested by each F. hybrid was calculated over mid parent (MPH%) = [(F1-MP)/MP x 100] and better parent, (BPH%) = [(F1-BP/BP x 100], respectively.

Observations on leaf rust incidence were recorded during peak periods of disease expression i.e., May- June months during pre-monsoon season and Sept.- Oct. months during post-monsoon season, for four successive years, from 2017 to 2020. Data on CLR incidence was recorded on individual plants and plants with less number of pustules were also treated as susceptible to assess the population susceptible/ resistant in each progeny. The disease build-up on susceptible plants was also scored using the 0-9 scale of Eskes and Toma-Braghini (1981 cf Eskes 1989), where 0 = plants are free from the symptoms; 1= presence of one diseased branch. Likewise, grades were assigned based on progression of disease and 9= maximum disease incidence. Finally, the plants were grouped into four categories viz., 1= tolerant (free from CLR incidence), 2 = moderately tolerant (mild infection without any defoliation); 3= susceptible (medium levels of incidence) and 4= highly susceptible (high disease build up coupled with defoliation).

All the four F. hybrid genotypes evaluated in the present study exhibited vigorous vegetative growth with compact bush stature which is expected when both the parents are semi-dwarfs. The data in respect of vegetative, yield and yield component characters is presented in Table 3.

Table 3
Character means and analysis of variance for different morphological and yield component characters among different F1 hybrids

From the data it is apparent that the hybrids exhibited uniform vegetative vigour with marginal differences in various characters. Among the F. progenies, S.5084 recorded superior growth in terms of stem girth (39.5 cm), number of nodes (20.2), bush spread (366.1); number of fruits per node (14.9). In contrast, S.5086 recorded more compact growth pattern as reflected in most of the growth parameters recorded, compared to other F. hybrid progenies (Table 3). Analysis of variance carried out for each character revealed that the character means among different hybrids are significant at P<0.05 with their parent populations except number of nodes per primary. These variations in terms of growth characters can be attributed to the growth pattern of the parents and their cross combination used for generation of hybrids. Prakash et al (2006) evaluated 17 elite hybrid progenies for growth parameters, field tolerance to rust, clean coffee yield and bean characteristics in a specific agro- climate in India. Analysis of variance of five morphological characters revealed significant differences between the genotypes in respect of four characters (stem girth, bush diameter, length of longest primary and number of primaries per plant) establishing that the genotypes are moderately heterogeneous for plant architecture. Heritability was high (60%) for yield and 64% for field tolerance to leaf rust indicating low genotype x environment interaction for these traits.

Data on year wise clean coffee yield among the F. progenies and the parents is presented in Table 4. In general, the production trend over the years from 2017-18 to 2020-21 has shown alternate bearing pattern which is common in Arabica coffee. Further, there has been an increase in quantum of yield both in parents and hybrids over the years. All the hybrids recorded higher yields over parents both during the on and off years. Statistical analysis of year wise yields showed significant differences among different genotypes and also the parents (Table 4). During the lean cropping years of 2017-18 and 2019-20, the mean crop yields (clean coffee) among the four hybrid progenies varied from 996 kg/ha (S.5083) to 1102 kg/ ha (S.5086) and 1161 kg/ha (S.5085) to 1232 kg/ha (S.5086), respectively. During the high cropping years i.e., 2018-19 and 2020-21, the mean yields ranged from 1233 kg/ha (S.5084) to 1424 kg/ha (S.5083) and 1415 kg/ha (S.5084) to 1611 kg/ha (S.5086). Among the hybrids, S.5086 recorded superior and consistent performance during individual years with a maximum yield of 1611 kg/ha during 2020-21 and the four year mean yield of 1313 kg/ha.

Among the F. hybrids, the relative heterosis (MPH%) ranged from 15.83% (S.5084) to 29.10% (S.5086) while heterobeltiosis ranged from 5.08% (S.5084) to 17.12% (S.5086). Among the four F. hybrids, S.5086 recorded superior yield performance.

Among the two parental lines, the yield in ‘Chandragiri’ ranged from 887 kg/ha (2017-18) to 1303kg/ha (2020-21) while in Sln.10, the year wise yields ranged from 741 kg/ha to 1041 kg/ha. Analysis

Table 4
Year wise clean coffee yields in F1 hybrids and parental lines

of relative heterosis and heterobeltiosis revealed that all the hybrids exhibited maximum relative heterosis (MPH%) that ranged from 20.7% (S.5084) to 37.5% (S.5086) during the high cropping year, 2020-21 except in S.5085. As regards to heterobeltiosis, all the hybrids except S.5083 recorded high BPH% that ranged from 10.26% (S.5084) to 24.24% (S.5086) during the low cropping year, 2017-18. These differences in relative heterosis and heterobeltiosis, could be attributed to the yield of the parents during the respective years. Among the four F. hybrids, S.5086 recorded high BPH% in all the years that ranged from 15.14% (2019-20) to 23.64 (2017-18) except during 2018-19.

Bertrand et al. (2005) evaluated the performance of F₁ hybrid plants derived from C. arabica for production variables and reported that the F₁ hybrids produced between 22% (trial 1) and 47% (trial 2) more fresh berries than the parental lines in two separate trials. This difference was highly significant for trial 2 (P = 0.00). From the studies on genetic parameters of Timor hybrid derived Arabica genotypes at IAC, Brazil, Mistro et al. (2007) reported that the greatest yield gains were achieved when selection was performed based on plot means and years of high yields. It was reported that under normal climate conditions, coffee yields usually increase from the first until the fourth/fifth year. Thereafter, the biennial yield cycles begin, characterized by the alternate high and low yields. From the results of the present study, it is apparent that both the parental lines and the hybrids reflected the alternate bearing behaviour. However, the consistency in production has been recorded in the hybrids in corresponding on and off years of production, alternatively.

Dula (2019) reviewed the heterosis and combining ability studies for yield of Coffea arabica varieties in Ethiopia. From the studies conducted by Mesfin and Bayetta (1983), the extent of heterosis for yield was up to 60% over better parent. Out of nine F. hybrids, only one hybrid exhibited negative heterosis of -8%. The highest yielding hybrids Melko-CH2 and Ababuna showed 20% and 18% heterosis over the better parent respectively. Given the fact that the genetic distance and combing ability of the parental lines is critical for achieving the maximum extent of heterosis in F₁ hybrids, the heterobeltiosis to the extent of 23.64% over better parent and 37.5% over mid parent for yield in best performing F. hybrid (S.5086) in the present study is a significant point to consider for commercial exploitation.

Data on coffee leaf rust (CLR) incidence among the four F₁ hybrids and parental lines is furnished in Table 5. Among the hybrids, the leaf rust incidence ranged from nil (S.5086) to 51% (S.5083) while in the two parental lines, the CLR incidence ranged from 2.8% to 30% in Chandragiri and 22% to 73% in Sln.10 during different years of study. From the data in respect of individual F₁ hybrids, it is apparent that the two hybrid progenies, S.5083 and S.5084 recorded relatively high susceptibility as 46.1% and 50.9% of population manifested susceptibility during 2019-20, the year that recorded high rust flare up due to favourable weather conditions. The parental lines also recorded maximum susceptibility (30% in Chandragiri and 73% in Sln.10) during the high rust year (2019- 20). In contrast, the remaining two F₁ hybrid progenies (S.5085 and S.5086) maintained high levels of field tolerance as the entire population of these two hybrids are free from the rust incidence. The high rust incidence manifested during the 2019-20 could be attributed to the favourable weather conditions i.e., high rain fall coupled with maximum number of rainy days during July and August months as well as ideal temperatures (190C to 240C). The quantum of rainfall though higher during July and August 2018, the number of rainy days were low thereby the rust incidence was relatively low compared to 2019. However, inspite of the favourable weather conditions during 2018 and 2019, the disease build up levels were recorded to be very low, in the hybrids S.5083, S.5084 and parent variety ‘Chandragiri’, the disease build up was characterized by the non-sporulating necrotic spots (group 2), indicating the high levels of tolerance. Further, the high levels of field tolerance to leaf rust in F. hybrid progenies (S.5085 and S.5086) could be attributed to the integration of the SH3 gene in these hybrid populations as the Sln.10 parent used as SH3 donor is homozygous to SH3.

In fact, the very objective of the F. hybrid breeding programme was to pyramid the maximum number of SH genes for rust resistance in a proven commercial arabica genotype to ensure long lasting resistance to leaf rust. The resistance in Chandragiri is governed by the SH genes of tetraploid arabica origin (SH1,2,4,5) and SH genes (SH6, SH7, SH8, SH9) introgressed to C. arabica from diploid species, C. canephora (robusta coffee). The F₁ hybrids evaluated in the present study were developed with the primary aim of pyramiding of SH3 gene of C. liberica origin in order to improve durability of resistance in Chandragiri. Two plants of Sln10, homozygous and heterozygous to SH3 gene were consciously selected and used as donor parents (paternal) in crossing programme. Apparently, the two hybrid progenies, S.5083 and S.5084 were derived from the heterozygous donor while the other two hybrid progenies, S.5085 and S.5086 were developed from the homozygous donor plant. Thus, the variability for field tolerance to leaf rust in these four F₁ hybrid progenies, could be attributed to the differences in pyramiding/ integration of SH3 gene from heterozygous and homozygous donor plants. Analysis of the selected F₁ plants representing the susceptible and resistant plants in different hybrid progenies with SCAR marker linked to SH3 gene confirmed the presence of the SH3 gene (un-published data). This inference lend credence from the findings of Shigueoka et al. (2014) evaluated nine arabica coffee progenies with an objective to select high- yielding coffee progenies with resistance to coffee leaf rust for the state of Parana in Brazil. It was reported that the plant population of several genotypes derived from ‘Sarchimor’ and ‘Catucai’ were susceptible to coffee leaf rust and complete resistance was broken in several coffee plants of ‘Catucai’ germplasm. However, the interesting observation reported was that the genotype ‘F₁ of Catuai x (Catuai x BA-10 coffee)’ resistance in more than 80% plants and inferred that the genotypes were heterozygous to SH3 gene. Infact, there were several other earlier reports from Brazil that the coffee genotypes carrying SH3 gene manifested complete resistance to leaf rust (Fazuoli et al., 2005, Pereira et al., 2005. Sera et al., 2007). The findings of the present study also in conformity with that of the earlier reports on high levels of field tolerance manifested by the two F. hybrids, S.5086 & S.5085 derived from crosses employing Sln10 plant homozygous to S 3 gene. probably a carrier of SH3 gene manifested complete H

Table 5
Leaf rust incidence % population in different F1 hybrids