Both the Ethernet II and IEEE 802.3 standards define the minimum frame size as 64 bytes and the maximum as 1518 bytes. This includes all bytes from the Destination MAC Address field through the Frame Check Sequence (FCS) field. The Preamble and Start Frame Delimiter fields are not included when describing the size of a frame.

Any frame less than 64 bytes in length is considered a "collision fragment" or "runt frame" and is automatically discarded by receiving stations.

The IEEE 802.3ac standard, released in 1998, extended the maximum allowable frame size to 1522 bytes. The frame size was increased to accommodate a technology called Virtual Local Area Network (VLAN). VLANs are created within a switched network and will be presented in a later course. Also, many quality of service (QoS) technologies leverage the User Priority field to implement various levels of service, such as priority service for voice traffic. The figure displays the fields contained in the 802.1Q VLAN tag.

If the size of a transmitted frame is less than the minimum or greater than the maximum, the receiving device drops the frame. Dropped frames are likely to be the result of collisions or other unwanted signals and are therefore considered invalid.

At the data link layer the frame structure is nearly identical. At the physical layer different versions of Ethernet vary in their method for detecting and placing data on the media.