Can anybody who also has the 1989 weigh it and verify if it should be KM# 10a in stead of KM# 10?
If your coins are in flips, no problem: weigh a 1 cent before 1989 and one after 1992 and see where the 1989 fits (1990 and 1991 are in both types, but probably only one exists. In Krause there are in only one type, but the wrong one, so the 1989 probably is in the wrong one too, and I'm afraid we took this mistake from there in our Numista catalog).
Maybe yet, maybe not. If everybody have the same weigh on their 1987 to 1991, they began the light one in 1987 and stop the heavy one in 1986/1987. If we have some cent weigh 3.1g and some 2.5g, like near 50/50, they produce the two varieties at the same time. The last option it's, if we are only a few people to have cent with a weigh of 2.5g, this is maybe a planchet error. Like during this year, the Royal mint struck the Bahamas cent in copper plated zinc with a weigh of 2.5g and a diameter of 19mm. The problem for me is the thickness. The Bahamas cent have a thickness on 1.41mm and mine Barbados 1987 cent have 1.5mm like the km#10b (light).
To put more mysteries about this, I found some unconfirm information that said tha the RCM (Royal Canadian Mint) struck 5,000,000 Barbados cent in 1987. In Numista (and any other participative site), we have that all of the 10,000,000 cent struck in 1987 was stuck by the Royal Mint in Llantrisant (UK).
Everywhere they said that al the coins of Barbados was struck in UK or by Franklin (USA) and the RCM on their web site said that they work with Barbados since 1978.
The information about Barbados coinage like to be incomplete, so we have the good chance the the light version began to be strike well before 1992.