question when reading note 139436.1 (cache fusion)

Hi All,

When I read the note 139436.1 in metalink, the following statement made me confused.

{code}
Let's assume there is a 3 node RAC cluster and with lock element 123 covering a
block in the EMP table:

User C issues a select against table EMP lock element 123 on Instance 3 which opens
an SL0 lock on Instance 3:

    ----------------      -----------------     ----------------- 
    |  Instance 1   |     | Instance 2    |    |  Instance 3    |
    |               |     |               |    |                |
    |   Lock Held   |     |   Lock Held:  |    |   Lock Held:   |
    | on LENUM 123: |     | on LENUM 123: |    | on LENUM 123:  |
    |               |     |               |    |      SL0       |
    |               |     |               |    |                |
    ----------------      -----------------     ----------------- 

{code}

As far as I know, no lock will be incurred for query. Why does this note say one shared lock will be opened?