Since we usually only get to see low-order M67s...

Assembly set for advanced collectors. (Image source: http://www.dtic.mil/dtic/tr/fulltext/u2/a500888.pdf)

This picture shows almost proper fragmentation. Desirably each blunt cone should form a single fragment with an irregular jagged base, roughly 2...3 mm in diameter and an average weight of around 100 miligrams to ensure a high hit probability. Note the distance between each line is 1 mm (Image source: https://projects.nfstc.org/trace/2011/presentations/Flohr-Different-Kind.pdf)

Mini arena test (with an open top) to collect the fragments from an M67. This test set-up was used if a a new detonator would suffice to high order detonate an M67. (Image source: https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2008/fuze/VACheng.pdf)

Dents caused by the high velocity fragments in the heavy gauge aluminium(?) sheets. (Image source: https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2008/fuze/VACheng.pdf)

1312 fragments were collected from the M67 with a standard C70 detonator. Since there was no roof installed we need to calculate the total amount of fragments: (1312/5)*6 = 1574,4. (Image source: https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2008/fuze/VACheng.pdf)
The M67 body w/o explosive weighs around 170g. That gives us 170g/0.1g = 1700 fragments nominally.

Fragmentation pattern of a M42/M46 Grenade Body (submunition). The M67 uses the same type of pattern to ensure correct fragment size and mass.
Edit: I have screwed up the thread title.

Assembly set for advanced collectors. (Image source: http://www.dtic.mil/dtic/tr/fulltext/u2/a500888.pdf)

This picture shows almost proper fragmentation. Desirably each blunt cone should form a single fragment with an irregular jagged base, roughly 2...3 mm in diameter and an average weight of around 100 miligrams to ensure a high hit probability. Note the distance between each line is 1 mm (Image source: https://projects.nfstc.org/trace/2011/presentations/Flohr-Different-Kind.pdf)

Mini arena test (with an open top) to collect the fragments from an M67. This test set-up was used if a a new detonator would suffice to high order detonate an M67. (Image source: https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2008/fuze/VACheng.pdf)

Dents caused by the high velocity fragments in the heavy gauge aluminium(?) sheets. (Image source: https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2008/fuze/VACheng.pdf)

1312 fragments were collected from the M67 with a standard C70 detonator. Since there was no roof installed we need to calculate the total amount of fragments: (1312/5)*6 = 1574,4. (Image source: https://ndiastorage.blob.core.usgovcloudapi.net/ndia/2008/fuze/VACheng.pdf)
The M67 body w/o explosive weighs around 170g. That gives us 170g/0.1g = 1700 fragments nominally.

Fragmentation pattern of a M42/M46 Grenade Body (submunition). The M67 uses the same type of pattern to ensure correct fragment size and mass.
Edit: I have screwed up the thread title.