Influence of spatial arrangement, biofertilizers and bioirrigation on the performance of legume – millet intercropping system in rainfed areas of southern India

In this study, we checked the potential of bioirrigation – defined as a process of hydraulic lift where transfer of water occurs from deep soil layers to top soil layers through plant roots. We tested this in a pigeon pea (PP) – finger millet (FM) intercropping system in a field study for two consecutive growing seasons (2016/17 and 2017/18) at two contrasting sites in Bengaluru and Kolli Hills, India. Our objective was also to optimize the spatial arrangement of the intercropped plants (2 PP:8 FM), using either a row-wise or a mosaic design. The field trial results clearly showed that spatial arrangement of component plants affected the yield in an intercropping system. The row-wise intercropping was more effective than mosaic treatments at the Bengaluru field site, while at Kolli Hills, both row-wise and mosaic treatment performed equally. Importantly, biofertilizer application enhanced the yield of intercropping and monoculture treatments. This effect was not influenced by the spatial arrangement of component plants and by the location of the field experiment. The yield advantage in intercropping was mainly due to the release of PP from interspecific competition. Despite a yield increase in intercropping treatments, we did not see a positive effect of intercropping or biofertilizer on water relations of FM, this further explains why PP dominated the competitive interaction, which resulted in yield advantage in intercropping. FM in intercropping had significantly lower leaf water potentials than in monoculture, likely due to strong interspecific competition for soil moisture in intercropping treatments. Our study indicates that identity plant species and spatial arrangement/density of neighbouring plant is essential for designing a bioirrigation based intercropping system.


59
Intercropping has been considered a sustainable way to utilize and share natural resources 60 among different crop species and to improve and stabilize crop yield (Brooker et al., 2015; 61 Martin-Guay et al., 2018). In intercropping systems two or more crop species are grown 62 together (Vandermeer, 1989). Crop yield in intercropping systems are often higher than in 63 sole cropping systems because resources such as soil moisture and nutrients are utilized more 64 efficiently (Dahmardeh et al., 2009;Lithourgidis et al., 2007;Martin-Guay et al., 2018). This 65 is because interspecific competition between intercropping partners is often lower than the 66 intraspecific competition so that a yield advantage occurs (Davis and Woolley, 1993). In 67 addition, beneficial effects of intercropping can come from resource facilitation. In rainfed areas of the arid and semiarid tropics, intercropping has also been suggested to 76 enhance the water availability of shallow-rooted crops via the facilitation of water by deep-77 rooted plants through hydraulic lift (HL) (Mao et al., 2012;Xu et al., 2008). The water 78 released from deep-rooted plants due to HL into topsoil layer becomes available to 79 neighbouring shallow-rooted plants, a process termed bioirrigation (Burgess, 2011). The 80 functionality of bioirrigation in intercropping systems has only been tested in a few studies -81 mainly under controlled conditions in the greenhouse. Sekiya and Yano (2002) showed in a 82 field experiment that pigeon pea (a deep-rooted legume) has the potential to perform HL and 83 could supply deep water to shallow-rooted maize. In another study, Sekiya et al. (2011) 84 showed that plants with deep roots are ideal for intercropping with shallow-rooted crops in 85 water limited agriculture fields and that this kind of intercropping system allows shallow-86  The facilitative and competitive interactions between PP and FM in response to the different 262 treatments were calculated using the LER. The LER indicates the efficacy of an intercropping 263 system for using natural resources compared with monoculture (Willey and Osiru, 1972 were packed into airtight Ziploc bags to avoid water loss; bags were kept in the dark and leaf 288 water potential was measured within 1 -2 hours after sampling. 289 290

Statistical analysis 291
Analysis of yield data and LWP from field trials was carried out using GraphPad Prism 292 software (version 7.0 for Mac OS X, GraphPad Software, La Jolla California USA). Data are 293 expressed as mean ± standard error of mean (SEM). Tukey`s test was used for post hoc 294 multiple treatment comparison following one-way ANOVA or multifactor ANOVA using 295 general linear models. The criterion for significance was p<0.05. 296

Total biomass, straw and grain yield per hectare and LER 298
Intercropping and the spatial arrangement of the intercropping partners had a significant 299 effect on the total biomass yield per hectare at the Bengaluru site in 2016-17 (Fig. 3, Table  300 2). In particular, the treatment T3+ produced significantly more biomass per hectare than 301 monocultures of the constitutive crops or other spatial arrangements at Bengaluru in 2016-17. 302 Likewise, treatment T3+ resulted in higher yields for straw and grain as compared to the 303 other treatments in 2016-17 at Bengaluru site (Fig. 3, Table 2). For the intercropping 304 treatments, total biomass yield, straw yield and grain yield all declined from the T3+ to T6+. 305 The results differed at the Kolli Hills site, where in 2016-17 PP (T2+) produced the highest 306 yields for total biomass, straw and grain and where FM (T1+) and the different intercropping 307 treatments produced slightly lower yields with no significant differences among each other 308 (Table 2). In summary, in 2016-17 we found a strong positive intercropping effect for total 309 biomass yield, straw yield and grain yield at Bengaluru site, where the intercropping effect 310 were strongest in the 8:2 row-wise spacing. In contrast, no yield improvements by 311 intercropping irrespective of the spatial arrangement were observed at the Kolli Hills site. 312 313 These observations are also reflected in LER values at Bengaluru site, where values for total 314 biomass were greater than one for T3+, T4+ and T5+ and where T3+ had the highest LER 315 value. Similarly for straw biomass, T3+ had higher LER values than T4+, T5+ and T6+. For 316 grain biomass LER values were greater than one for the T3+ and T4+ treatment, equal to one 317 for T5+ and less than one for T6+ ( Fig. 4). At Kolli Hills LER values for all treatments were 318 less than one (Fig. 4). 319 320 In 2017-18, intercropping and the spatial arrangement of the intercropping partners also had a 321 strong and significant effect on the total biomass yield, straw yield and grain yield at 322 Bengaluru site (Fig. 5). As in 2016-17 the treatment T3-and T3+ produced significantly 323 more biomass per hectare than monocultures of the constitutive crops or other spatial 324 arrangements when compared to the respective treatments with and without biofertilizer. 325 Importantly, the application of biofertilizers enhanced the total biomass yield, straw yield and 326 grain yield in all treatments and this effect was consistent irrespective of experiment site, 327 mono or intercropping (Table 3). At Kolli Hills, we also found significant treatment effects 328 (Fig. 5). However, intercropping treatments did not produce higher yields for total biomass 12 and straw than any of the other treatments with or without biofertilizer. Yet, treatment T5+ 330 was equal in total biomass yield than the most productive monoculture (T2+). For grain yield 331 FM monoculture exceeded the productivity of PP (Fig. 5f) and in intercropping T3-, T3+ and 332 T5+ grain yield was similar to monoculture of FM with or without biofertilizer. The effects 333 of biofertilizers on total biomass yield, straw yield and grain yield that we detected at the 334 Bengaluru site were also observed at the Kolli Hills site and this effect was again consistent 335 across all treatments (Fig. 5, Table 3). We did not find a significant interaction between 336 treatment and biofertilizers nor a significant three way interaction between treatment, 337 biofertilizers, and site. However, as indicated above, the effects of biofertilizers at Kolli Hills 338 resulted in total biomass yield, straw yield and grain yield that were of the same magnitude in 339 some intercropping treatments as the highest yield in the corresponding monocultures (e.g. 340 T5+ for total biomass yield, and straw yield, and T3+ and T5+ for grain yield) (Fig. 5). In 341 summary, in 2017-18 we found a strong positive intercropping effect for total biomass yield, 342 straw yield and grain yield at Bengaluru site. In Kolli Hills, no such intercropping effect was 343 found. Importantly, biofertilizers improved the yields of crops in both sites and independently 344 of treatment. Despite the nonsignificant biofertilization -treatment interaction, intercropping 345 treatments at Kolli Hills showed yet a trend to be more enhanced through biofertilizers than 346 monocultures to an extent that they produced similar yields than the most productive 347 monoculture, which we did not observe without biofertilizers. in contrast, total biomass per plant was largest in treatments T4+ and T5+ compared to T2-378 and T2+ (Fig. 8d). 379 380 A two-way ANOVA analysis was performed to test the effects of spatial arrangement and 381 biofertilization on per plant yield (Table 8 &  Effects of the spatial arrangements can be explained by intra-and interspecific competition, 474 as illustrated when data are expressed per plant biomass (Fig. 7 & 8). Results from Bengaluru 475 clearly indicate that PP benefits in terms of per plant biomass in intercropping treatments 476 likely due to reduction in intra-specific competition that PP faces in monoculture. In contrast, 477 ; therefore, spatial arrangement between the plants 516 needs to be carefully optimized. In this study, PP had a head start of 45 days (polybag 517 transplantation) compared to FM, which provided PP a competitive advantage to acquire 518 more resources (light, nutrients and water) through its well-established root network, and FM 519 may face, additionally, shading effect due to tall PP plants. 520 521

Effects of biofertilizers 522
In the 2017-18 field trial, at both experimental sites, the effect of biofertilizer application was 523 positive and showed an increase in total yield (Fig. 8). The positive effect of biofertilization 524 did not differ among intercropping treatments with different spatial arrangements (Table 8 & 525 9). The effect of biofertilization was, however, specific to each component plants in the PP-526 FM intercropping system. Total biomass and straw yield per plant in FM was not 527 significantly affected by biofertilization, but grain yield was significantly increased (Table 6)

Effect of intercropping and biofertilizers on water relations of FM 543
In this study, the water relations (predawn LWP) of FM decreased significantly in mosaic 544 treatments as compared to row-wise and monoculture treatments (Fig. 9a & 9b). The trend in 545 predawn LWP (Fig. 9a & 9b) can also be compared with the trend in biomass production per 546 plant (Fig. 7a & 8a), therefore, competition for water could be the limiting factor here which 547 influenced the yield and effectiveness of intercropping treatments at Bengaluru site. Our 548 results suggest that there exists an important degree of below-ground competition for water 549 between PP and FM, and the facilitative effect of bioirrigation is suppressed. Similar results 550 have been reported by Ludwig et al. (2004). They found that HL performing trees extracted a 551 significant amount of water from the topsoil layer that resulted in lower LWP in understorey 552 grasses; however, grasses were able to absorb soil moisture released by tree due to HL. 553 554 One of our objectives was to find out if CMN can facilitate the transfer of bioirrigated water 555 from PP to FM and improve the water-relations of FM in intercropping treatments. The 556 results from the 2017-18 field trial showed that CMN did not affect the water relations 557 (predawn LWP) of FM in intercropping treatments. However, at week 1 and 2 (first and 558 second week of November 2017) FM in T3+ had higher, but not significant, LWP than T3-. 559 Similarly, FM in monoculture treatment showed a higher (less negative LWP) with CMN 560 than without CMN (Fig. 9b). Since, we observed similar effects of CMN in both monoculture 561 and 2:8 row-wise intercropping, we cannot assign this to bioirrigation. The effect of CMN 562 changed over time, and at week 3 (third week of November 2017) treatments T1+, T3+, and 563 T5+ (with CMN) had a lower LWP than T1-, T3-, and T5-(without CMN). The effect of 564 different treatments, biofertilization and times (weekly measurement) had significant 565 interaction with each other (Table 10). In this study, we showed that intercropping has a positive effect on total yield of PP and FM 577 but this effect varies across the sites based on site characteristics such as soil type and 578 weather. In conclusion, the answers to our three research questions are as follows: (i) the 579 spatial arrangement of intercropping partners does affect the straw and grain yield in a FM -580 PP intercropping system, and the optimal spatial arrangement for PP -FM intercropping 581 system depends on geographic location (local weather conditions) and plant variety. In 582 general, the row-wise treatment (T3+) resulted in better yields than the mosaic treatments at 583       (Fig. 3a, 3c & 3e) and Kolli Hills (Fig. 3b, 3d & 3f) during year 2016/17. Bars represent the average of four replicates with standard error of mean. One-Way ANOVA followed by Tukey`s test (posthoc test) was used for the combined biomass of FM and PP, separately for each site, and values with same letters are not significantly different from each other at p>0.05.