We test the microstructure invariance proposed by Kyle and Obizhaeva (2016) in the grain markets. Using the CME’s intraday best-bid-offer data from 2008 to 2015, we find support for both trade size invariance and trading cost invariance at 1-minute, 5-minute, and 10-minute, although not in its original form. After rescaling the trading activity by spread cost per Benzaquen et al (2016), we find strong evidence for both hypotheses of invariance. The findings help understand the trading dynamics of grain commodities from both trading and regulatory perspectives. Specifically, we can derive the number of trades, trading cost, and illiquidity measure based on observable metrics, such as price, volume and historical volatility. These imputed measures can be further used to identify the systematic risks resulting from speculative transactions.