The following assumes that AGE is the only explanatory variable; 
another possibility would be AGE and WHEEZE as explanatory, e.g...

402 > coal _ fac.design(c(2,2,9),c("Wheeze","Breath","Age"))
402 > coal$Resp _ scan("t43.dat")
402 > coal.glm1 _ glm(Resp ~ Age*Breath*Wheeze,data=coal,family=poisson)
402 > qqnorm.aov(coal.glm1)
402 > args(qqnorm.aov)
function(fit, full = F, label = F, omit = NULL, xlab = paste(if(full) "Full-"
	 else "Half-", "normal Quantiles", sep = ""), ylab = "Effects", ...)
NULL
402 > qqnorm.aov(coal.glm1,label=6)
402 > coal.glm2 _ glm(Resp ~ Age + Breath*Wheeze,data=coal,family=poisson)
402 > anova(coal.glm2)
Analysis of Deviance Table

Poisson model

Response: Resp

Terms added sequentially (first to last)
              Df Deviance Resid. Df Resid. Dev 
         NULL                    35   25889.47
          Age  8   886.64        27   25002.83
       Breath  1 11026.22        26   13976.61
       Wheeze  1  7037.54        25    6939.07
Breath:Wheeze  1  4237.13        24    2701.94

402 > attach(coal)
402 > plot.factor(Wheeze,log(Resp))
402 > plot.factor(Breath,log(Resp))
402 > plot.factor(Age,log(Resp))
402 > interaction.plot(Breath,Wheeze,log(Resp))
402 > interaction.plot(Age,Wheeze,log(Resp))
402 > interaction.plot(Age,Breath,log(Resp))
402 > detach()
402 > coal.glm3 _ glm(Resp ~ (Age + Breath + Wheeze)^2,data=coal,family=poisson)
402 > anova(coal.glm3)
Analysis of Deviance Table

Poisson model

Response: Resp

Terms added sequentially (first to last)
              Df Deviance Resid. Df Resid. Dev 
         NULL                    35   25889.47
          Age  8   886.64        27   25002.83
       Breath  1 11026.22        26   13976.61
       Wheeze  1  7037.54        25    6939.07
   Age:Breath  8  2342.40        17    4596.68
   Age:Wheeze  8  1544.82         9    3051.86
Breath:Wheeze  1  3025.17         8      26.69
402 > pchisq(26.69,8)
[1] 0.9992004