We have identified 21,772 Child Health and Development Studies grandchildren (F2) by linking to the California Birth Records, which was slightly greater than our original estimate of 20,000. Using address history files, we verified 18,371 CHDS grandchildren births. By linking the verified births to the California Department of Developmental Services records we have successfully identified 116 autism cases in our cohort which exceeds our initial estimate of 72. We found Grandpaternal smoking was related to ASD (adjusted RR=2.03, 95% confidence interval [1.25, 3.3]), but grandmaternal smoking was not (aRR=0.9 [0.6, 1.5]). Increasing total alcohol consumption was suggestive of increased ASD risk (aRR=1.04 [0.96, 1.13] per grandmaternal drink consumed; aRR=1.04 [0.99, 1.10] per grandpaternal drink consumed), with stronger associations for hard alcohol (aRR=1.11 [1.00, 1.24] per grandpaternal drink consumed; aRR=1.15 [0.97, 1.36] per grandmaternal drink consumed). There was no association of ASD with increasing wine or beer consumption, or with any vs. no alcohol consumption, for either grandparent. Adjusting for F1 covariates including growth and development did not alter results. This is the first study of its kind in the United States, linking three generations studying germline exposures and risk of autism.