estimability

I have the following data:

E B T Y
2 1 1 11
2 1 2 15
2 1 3 18
2 1 4 25
2 2 1 36
2 2 2 52
2 2 3 41
2 2 4 14
1 1 1 15
1 1 2 25
1 1 3 25
1 1 4 29
1 2 1 65
1 2 2 45
1 2 3 23
1 2 4 32
1 3 1 54
1 3 2 14
1 3 3 15
1 3 4 25

the model

Prueba
  e 2
  b 3
  t 4
  y
aaa.txt !skip 1
y ~ mu e at(e).b !r t t.e
predict t.e

and this is the .pvs file

  Prueba 31 Jan
2008 14:04:29
                                                             MYT\PruebaAsreml\aaa

  Ecode is E for Estimable, * for Not Estimable

  ---- ---- ---- ---- ---- ---- ---- ---- 1 ---- ---- ---- ---- ----
---- ---- ----
  Predicted values of y
  b is averaged over fixed levels
  The cells of the hypertable are calculated from all model terms
constructed
                 solely from factors in the averaging and classify
sets.
  Warning: 4 non-estimable [aliased] cell(s) may be omitted from
the table.
                 The Overall SED statistic includes non-estimable
predictions.

  t e Predicted_Value Standard_Error Ecode
        1.0000 1.0000 31.5336 4.4375 E
        2.0000 1.0000 30.4090 4.4375 E
        3.0000 1.0000 29.9367 4.4375 E
        4.0000 1.0000 30.4540 4.4375 E
  SED: Overall Standard Error of Difference not available

Why the blups for e 2 are non-estimable?

Thanks
Received on Sun Jan 31 2008 - 14:07:53 EST