LimitPredicate

Hallo,

I am having trouble detecting specific nodes in a node-set that appear on a
standard frequency.
For example, i have the following node-set which always comprises of a
number of ITEM nodes multiple to
three.So i can have 3 or 6 or 9  etc, but i don't know the exact number
each time because this xml
is produced dynamically.In this particular example i have exactly 15 ITEM
nodes.

<ITEMS>
<ITEM>1</ITEM>
<ITEM>2</ITEM>
<ITEM>5</ITEM>
<ITEM>7</ITEM>
<ITEM>2</ITEM>
<ITEM>5</ITEM>
<ITEM>10</ITEM>
<ITEM>2</ITEM>
<ITEM>5</ITEM>
<ITEM>15</ITEM>
<ITEM>2</ITEM>
<ITEM>5</ITEM>
<ITEM>17</ITEM>
<ITEM>2</ITEM>
<ITEM>5</ITEM>
</ITEMS>

Now if i want to select the values of ITEM nodes that appear for example in
the 1st, 4th, 7th etch
position, i use select="ITEM[position()mod 3=1].Using this predicate i get
to select the
numbers 1,7,10,15 and 17.What if i wanted to limit my selection and skip
the third selected value
,that is 10, and get to select all the others.How do i do this, i mean
select all ITEM contents
whose position is denoted by mod 3 =1, except for the third of these
particular items whose content
is 10.Is there a way to further limit the predicate syntax to achieve
this.A similar task would be
to just select the fourth of the nodes that their position is like mod 3
=1, and simply select
the value 15 in this particular example.How can i declare this with XPath.I
mean
can i write something like select"ITEM[position() mod 3 = 1](4) to select
just the fourth node
among the particular nodes.I know that what i just wrote is incorrect, but
how can this
be done correctly??

Regards
Manousos
Athens