and then suggested at the end:
which is paired with the bass line:
But in these last four measures, the other voices do not follow the pattern of the sequence. After singing through each part, I realized that the potential of certain lines, especially the tenor seems lost. The three C's create a suspension-type pattern, which when paired with passing tones in the other voices, stresses beats 1 and 3:
This powerful combination sets up the expectation to be used again in sequence with the soprano and bass, but this never happens. Instead, the tenor line seems rather odd: (C->C->C-Bb-A, G-C-C-A) followed by more repeated C's in the last bar (C-C-C-Bb, A), which now don't imply direction, but rather filler for the harmonic progression. Furthermore, the harmonic progression uses a borrowed V/ii, which is suggested in measure two, but then alto line moves to an E, suggesting a diminished vii in first inversion.
Although the meter of the text does suggest two-part phrases (18.104.22.168 D), many of the sentences in the text do not have punctuation at this moment (e.g. "though all the mighty billows shake the mountains on the shore" [vs. 1] and "for God will hasten to her aid when trouble is at hand" [vs.2] ) The first line "God is our refuge and our strength, our ever-present aid" is a nice two-part, punctuated phrase, but if this idea is continued for every line in double meter, the whole hymn can become a bit disjointed and bland. Adding connective material, like in the last phrase would smooth out the flow of the melodies.
So, I began to rewrite. Should be easy enough to continue the tenor sequence:
The result when paired with soprano and bass does have pretty biting dissonances, m9 and M7 (tenor Bb vs. two A's in bass and soprano), but because it is a more pure sequence, the direction seems natural and the resolutions provide a certain and tug-and-pull that wasn't in the original second measure. The alto line moves to D instead of E to further suggest Gm (ii) and the third measure continues the harmonic sequence to F major in root position, not first inversion. It seemed natural to add an alto 4-3 suspension, which sets up a chain of eight-note passing tones which is echoed by the tenor.
Sullivan's original part-writing is correct and does not sound unpleasant, but to me it feels like the potential of this last phrase was lost. By further suggesting the sequential nature of the soprano and bass, giving more time to colorful harmonic progressions (V/ii --> ii), and allowing each voice to share in the passing tones and suspension more equally, this final phrase seems more complete, with more direction and balance between turbulence and resolution.
But, then the problems began. My first attempt yield parallel octaves between soprano and tenor (m.1 into m. 2) If the tenor starting note of the sequence is the root of the chord, not the third, the sequence can resolve down to the root of the next sequence level. I also considered how the alto motion, leaping from G-E in the original, can be used in both statements of the sequence. In the third version, I mainly am experimenting with a different bass line, which uses more passing tones, which allows for the final low F to be saved until the end, but this results in another parallel octave between bass and alto before measure 3.
The final measure still has some issues with parallel octaves in tenor and bass, present in all of my versions, which are disguised by passing tones. If you keep the original voicing of first inversion, one of the them is relieved (F-A), but other remains (A-C).
Green - sequential progression
Blue - passing tones / neighbor tones
Orange - suspensions
Red - part writing that concerns me