The baryonic Tully–Fisher relation ( BTFR ) is both a valuable observational tool and a critical test of galaxy formation theory .
We explore the systematic uncertainty in the slope and the scatter of the observed line width BTFR utilizing homogeneously measured , unresolved H i observations for 930 isolated galaxies .
We measure a fiducial relation of \log _ { 10 } { M _ { baryon } } = 3.24 \log _ { 10 } { V _ { rot } } ~ { } + ~ { } 3.21 with observed scatter of 0.25 dex over a baryonic mass range of 10 ^ { 7.4 } to 10 ^ { 11.3 } M _ { \odot } where V _ { rot } is measured from 20 % H i line widths .
We then conservatively vary the definitions of M _ { baryon } and V _ { rot } , the sample definition and the linear fitting algorithm .
We obtain slopes ranging from 2.64 to 3.53 and scatter measurements ranging from 0.14 to 0.41 dex , indicating a significant systematic uncertainty of 0.25 in the BTFR slope derived from unresolved H i line widths .
We next compare our fiducial slope to literature measurements , where reported slopes range from 3.0 to 4.3 and scatter is either unmeasured , immeasurable or as large as 0.4 dex .
Measurements derived from unresolved H i line widths tend to produce slopes of 3.3 , while measurements derived strictly from resolved asymptotic rotation velocities tend to produce slopes of 3.9 .
The single largest factor affecting the BTFR slope is the definition of rotation velocity .
The sample definition , the mass range and the linear fitting algorithm also significantly affect the measured BTFR .
We find that galaxies in our sample with V _ { rot } < 100 km s ^ { -1 } are consistent with the line width BTFR of more massive galaxies , but these galaxies drive most of the observed scatter .
It is critical when comparing predictions to an observed BTFR that the rotation velocity definition , the sample selection and the fitting algorithm are similarly defined .
We recommend direct statistical comparisons between data sets with commensurable properties as opposed to simply comparing BTFR power-law fits .