Why do certain new crop varieties fail to gain popularity among farmers despite their proven superiority in scientific field trials and on-farm demonstrations? Empirical microlevel studies in agricultural economics usually address this question from a household or farmer perspective, neglecting the broader socio-institutional and economic context against which varietal dissemination takes place. In the present study, we explore the relationship between wheat seed delivery networks and varietal turnover, using data from farmers' fields in eastern India. An extensive survey encompassing 7648 wheat farmers was conducted to capture wheat production practices across 31 districts of Bihar and 9 districts of eastern Uttar Pradesh. Additional qualitative interviews led to the initial impression that farmers' selection of varieties often precedes their choice of seed source, challenging the prevailing assumptions about the causal direction of this relationship. To address the challenge of reverse causality between varietal selection and seed source preference, we employed an instrumental variable regression framework. Although new wheat varieties are available in the market, a substantial majority of farmers opt for old ones. Farmers' decisions regarding the selection of wheat varieties and seed sources are not straightforward, and their preference for wheat varieties significantly impacts their seed source selection, indicating a hitherto unexplored bidirectional relationship. Understanding this complex relationship has significant strategic implications for stakeholders in the seed supply system, offering policy recommendations for enhancing varietal turnover rates and encouraging the widespread adoption of new, improved varieties among smallholders. Such initiatives hold the promise of enhancing system resilience and food security in developing countries.