First in AMOS can use as data input a correlation matrix . Here is a link (download the .sps syntax file)
http://core.ecu.edu/psyc/wuenschk/MV/SEM/Path-SPSS-AMOS.doc and,
http://groups.google.com/group/c ... 9f08cf7766d2b?pli=1

Second, if your reliability is low, usually<0.6 that means the item correlations are also low. That is you can see in SPSS (check all the descriptives under the statistics button) in reliability dialog. So, you can drop the problematic variables. Furthermore, you must see if any of your items needs reverse coding according to your questinnaire. After that when your reliability is acceptable, you can import your correlation matrix.

Third there are some rules how many indicators are in any latent variable. The next is from Kline's book(page 138), "PRINCIPLES AND PRACTICE OF STRuCTuRAL EQuATION MODELING, Third Edition
Rex B. Kline".

If a standard CFA model with a single factor has at least three (Rule 6.4)
indicators, the model is identified.
If a standard CFA model with ≥ 2 factors has ≥ 2 indicators per (Rule 6.5)
factor, the model is identified.

I wish these are helps you