Melanie Epp Can grain be grown without synthetic inputs? This Serbian farm is testing it Samo Login sold his talking cat app for nearly US$1 billion, then invested the proceeds in a theory. The Slovenian entrepreneur wanted to know whether non-GM grain and legumes could be grown at commercial scale without synthetic inputs or manure. He had not set out to become a farmer. After selling Outfit7, he went looking for the most under-invested, highest-impact problem he could find and landed on food production. His reasoning centred on nitrogen and phosphorus pollution from agriculture, not carbon emissions. To answer his question, he bought an 8,100-acre farm in Serbia and hired Djura Karagić, a local agronomist, to turn that theory into practice. Karagić did not arrive as a believer. He came from conventional research, where he led the forage crops department at Serbia's National Research Institute in Novi Sad. By his own account, he joined LoginEKO with scepticism, the way most of his colleagues would have. The neighbours were certain he would fail, and, for the first two seasons, they LoginEKO is testing whether grain can be grown at commercial scale with lower input costs, but the model depends heavily on local conditions, machinery, and capital. ONTARIO GRAIN FARMER 34 AGRONOMY were right. Borrowed techniques that were successful in other climates and contexts did not translate well in heavy clay soil under increasingly erratic rainfall. Weeds, at one point, grew taller than the farm's tractors, he said. Karagić called this period the "YouTube agriculture" phase. To make the theory work, he would need to develop a plan as unique as the environment. He did not have to do it alone. Karagić works alongside the LoginEKO team, which is made up of agronomists, field managers, machinery operators, and data and software colleagues, all of whom played an essential role in developing, testing, and improving the system. He also relied on guidance from researchers at Wageningen University & Research and Oxford University, who helped assess, validate, and independently evaluate the sustainability of the model. Most of what the LoginEKO team knows now is the result of mistakes made in those early seasons. Today, new ideas are tested first on small experimental plots, then verified on several larger fields across two locations. Crops are only scaled up when results hold. LoginEKO has trialled more than 50 crops this way, landing on five or six that fit the rotation. Wheat provides the clearest example of how much trial and error went into the process. Karagić said they tried nearly everything to raise protein content, including cover crops and green manure peas, before concluding that, in a semi-arid climate, a green manure crop simply cost them a season of commercial grain without delivering enough nitrogen to justify it. The breakthrough came from a variety screening trial. They tested 27 wheat varieties and identified two capable of reaching 15 per cent protein in good conditions. In practice, they are now consistently hitting 13.5 per cent — enough to clear the 12 per cent threshold to be sold as bread wheat instead of feed. The same trial-and-scale approach has been used to shape the legume side of the rotation. Field peas, the backbone of the farm's nitrogen supply, are well suited to this climate because they mature early enough to avoid the worst of the summer drought. Chickpeas, a new crop for this part of Serbia, are being tested using the same approach. Karagić is trialling four seeding dates and using a U.K.-sourced bacterial inoculant to build a second, more modest nitrogen source alongside the peas. Yield results are mixed. Field peas average 37 to 39 bu./ac. Chickpeas, still in the onfarm trial phase, are targeting about 30 bu./acs but are not there yet. The rotation itself runs in five-year cycles, which vary in terms of crops included. A typical version starts with winter peas to fix nitrogen, moves into a cereal — wheat or oats — works in chickpeas, then winterhardy flax, and closes with sunflower, a deep-rooted crop that scavenges
RkJQdWJsaXNoZXIy MTQzODE4