Nutrients, light and different applicable stresses are responsible.

usually, we can see some difference in germination rate and percent, emergency, etc in pure lines or varieties.

If you have higher variation at two sites than at a third, you should look to see what is different (presumably better) at the third.

• Are you certain that the seedlots are identical?
• Are seed storage and sowing practice also the same, perhaps inferred from germination rate.
• At the variable site consider variation (i) between sowings and (ii) within a sowing.
• If between sowings, batches of rooting medium may be relatively variable or seasonal effects are large.
• If within a sowing, ie. due to position effects in the propagation environment or nursery, look for variation in watering, direct sun etc.

This is probably as much a management problem as a research problem.  

The variation was found due to seed size, in the same variety, different seeds can be found, due to well filled, less filled, partially filled etc. Other reasons due to nutrient, light, shading, biotic and abiotic stresses etc. This was found in Aman and Aus because the rice has the normal growth due to optimum temperature but in Boro (in Bangladesh) due to winter (cold stress), the normal growth of rice was paused and all seedlings were stunted, that's why you don't find any difference in the same variety. 

It would be unusual for root penetration into a (polystyrene?) tray to be advantageous to growth unless the roots were able to grow into a neighbouring cell or into the soil. The greater effect is more likely to be the damage caused to such a root system in removing the seedling from the tray, since those roots penetrating the tray would have to be broken. This problem can be overcome by coating the trays in a copper solution before filling with rooting medium. The copper inhibits root extension growth, makes the root plug more cohesive, may improve root regeneration potential and increases the life of the trays.