proc mixed - LMS question

[Ziyad]

Dear

I was running proc mixed to my data

Stallion	Ejaculate	Treatment		Bacillus
1	1	CW		0
1	1	CA		0
1	1	SW		0
1	1	SA		0
2	1	CW		0
2	1	CA		0
2	1	SW		0
2	1	SA		0
3	1	CW		0
3	1	CA		0
3	1	SW		1000
3	1	SA		1000
1	2	CW		0
1	2	CA		0
1	2	SW		10000
1	2	SA		1000
2	2	CW		0
2	2	CA		0
2	2	SW		0
2	2	SA		0
3	2	CW		0
3	2	CA		0
3	2	SW		0
3	2	SA		0
1	3	CW		0
1	3	CA		0
1	3	SW		0
1	3	SA		0
2	3	CW		0
2	3	CA		0
2	3	SW		0
2	3	SA		0
3	3	CW		0
3	3	CA		0
3	3	SW		0
3	3	SA		0
4	1	CW		0
4	1	CA		0
4	1	SW		0
4	1	SA		0
5	1	CW		0
5	1	CA		0
5	1	SW		0
5	1	SA		0
6	1	CW		0
6	1	CA		0
6	1	SW		0
6	1	SA		0
4	2	CW		0
4	2	CA		0
4	2	SW		0
4	2	SA		0
5	2	CW		0
5	2	CA		0
5	2	SW		0
5	2	SA		0
6	2	CW		0
6	2	CA		0
6	2	SW		0
6	2	SA		0
4	3	CW		0
4	3	CA		0
4	3	SW		0
4	3	SA		0
5	3	CW		0
5	3	CA		0
5	3	SW		0
5	3	SA		0
6	3	CW		0
6	3	CA		0
6	3	SW		0
6	3	SA		0

by used  this code

TITLE "bacillus";
PROC MIXED DATA=Bacteria2 PLOTS=PearsonPanel;
	class stallion Ejaculate  treatmeant;
	model bacillus = treatment;
	RANDOM stallion  stallion*Ejaculate stallion*treatmeant;
	LSMEANS treatment /adjust=Tukey;
RUN;

the problem in CA and Cw the mean is zero therfore the least Squares Means un readable as in picture

what is the least Squares Means her.

Best Regards.

[Rick_SAS]

I'm not sure what you mean by "unreadable." The bacillus variable is identically zero when Treatment='CA' or Treatment='CW'. The numbers being displayed for the LSmeans are tiny numbers on the order of machine roundoff. The numbers -227E-15 is scientific notation for -227 x 10^{-15}, which is -2.27 x 10^{-13}. Only the SW treatment has an LSMean that is significantly different from 0.

[Ziyad]

That is mean i use this number as LSmeans even if it in minus.