Unexpected results using Anova and TukeyHSD when I halve the groups

Hello,
I am new to R and hopeless with statistics and its for my PhD.
I am applying TukeyHSD to my simple problem but I get an answer I don't understand.

I have 6 groups of data, and I conduct an Anova test and get p =.45
when I do a TukeyHSD I get a minimum p=.88 on all the groups
If I split the groups into 2, the first group (2L) I get a p =.07
I would have expected the .07 to appear in the TukeyHSD 6 group test
so it's like I need to split into groups (2L and 5L) to get a better p-value but that doesn't make sense so I was hoping someone could clear give me an explanation.
Thanks

Code:
attach(R_solid)
aov(Artery~Emboli)
summary(aov(Artery~Emboli))
TukeyHSD(aov(Artery~Emboli))

R_solid
Emboli Artery
1mm (2L) 67
1mm (2L) 43
1mm (2L) 89
2mm (2L) 16
2mm (2L) 9
2mm (2L) 52
3mm (2L) 13
3mm (2L) 6
3mm (2L) 41
1mm (5L) 81
1mm (5L) 9
1mm (5L) 33
2mm (5L) 60
2mm (5L) 1
2mm (5L) 23
3mm (5L) 71
3mm (5L) 0
3mm (5L) 45

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.