We investigate the nature of objects in a complete sample of 28 galaxies selected from the first sky area fully covered by ALFALFA , being well-detected and having HI profiles wider than 550 km s ^ { -1 } .
The selection does not use brightness , morphology , or any other property derived from optical or other spectral bands .

We investigate the degree of isolation , the morphology , and other properties gathered or derived from open data bases and show that some objects have wide HI profiles probably because they are disturbed or are interacting , or might be confused in the ALFALFA beam .
We identify a sub-sample of 14 galaxies lacking immediate interacting neighbours and showing regular , symmetric , two-horned HI profiles that we propose as candidate high-mass disk systems ( CHMDs ) .

We measure the net-H \alpha emission from the CHMDs and combine this with public multispectral data to model the global star formation ( SF ) properties of each galaxy .
The H \alpha observations show SFRs not higher than a few M _ { \odot } yr ^ { -1 } .
Simple SF models indicate that the CHMDs formed most of their stars almost a Hubble time ago , but most also underwent an SF event in the last \sim 10 ^ { 6 - 7 } yrs ; the young stars now produce 10 to 30 % of the visible light .
The spatial distribution of the SF-regions is compatible with recycled stellar ejecta .

We calculate representative dynamical masses M _ { dyn } for the CHMDs ranging from \sim 10 ^ { 11 } M _ { \odot } to \sim 7 \times 10 ^ { 11 } M _ { \odot } .
The M _ { dyn } values are larger by factors of 2.5 to 7.5 than the baryonic masses of the luminous stars and gas but are affected by rather large errors .
We test the Tully-Fisher relation for the CHMDs and show that these lie below the relation defined by lower mass galaxies , i.e. , that their M _ { dyn } is lower than expected when extrapolating the relation from lower mass galaxies to higher HI line widths .