Series.apply bug when result same length as series

```
In [2]: s = pandas.Series([1,2])

In [3]: s
Out[3]: 
0    1
1    2

In [4]: s.apply(lambda x: (x, x+1))
Out[4]: 
0    0    1
1    2
1    0    2
1    3
```

But:

```
In [8]: s = pandas.Series([1,2,3])

In [9]: s
Out[9]: 
0    1
1    2
2    3

In [10]: s.apply(lambda x: (x, x+1))
Out[10]: 
0    (1, 2)
1    (2, 3)
2    (3, 4)
```
