The table calls it a tie: Russian Sage and Catmint share the crown at 86 and 86 of 100, a gap smaller than the data can honestly split.
A bee garden is judged by the calendar, not by a single July photo: the ranking pays most for months of documented nectar, then for surviving rough strip-of-lawn ground without care. Species that host specialized native bees get their own tier on top.
The picks
The flowers that won
The top score, plus the qualifier that asks least of you.
5 rules, 100 points, no opinions. Every line reads one field of the table.
Hosts specialized native bees10 pointsdocumented host of specialized native bees 10; general bee plant 4 Read from: pollinator groups.
Season-long nectar35 pointseight or more months 35; six to seven 31; four to five 27; three 17; two 9; one 4; no bloom data 9 Read from: the months it is in bloom.
Care while it grows20 pointslow upkeep 20; middling 11; demanding 3; sources disagree, we average the range; no rating 9 Read from: how much upkeep extension guides report.
Holds rough pollinator-strip ground25 pointsshrugs off both drought and poor soil 25; drought only 19; poor soil only 12; neither documented 6; no data 10 Read from: stress tolerances.
A pollinator strip deer walk past10 pointsRutgers A, rarely damaged 10; B 7; C 3; D, deer candy 0; not rated 5 Read from: the Rutgers deer rating.
Not ranked here: no whitelisted source documents bee visits.
The full method ↗
The leaderboard
The top ten
The 10 highest of 55 that qualified. Tap any score to open the lines behind it.
Not qualifying is a verdict too: the table simply says this question is not theirs.
no whitelisted source documents bee visitsTulips, Daffodils (Narcissus), Chrysanthemum and 27 more
The winner's badge
Generated by the ranking above, and re-scored whenever the table changes.
Russian Sage
The badge is generated by this ranking, not awarded by hand. When the
table changes, it re-scores itself.
How we rank
We keep one table of 85 flowers,
with 25 documented facts each, drawn from university and extension sources. Every ranking is
a published formula worth 100 points; tap any score and it unfolds into the lines that made
it. A flower that does not fit the question is not scored lower: it is excluded, and the reason is
shown. When the table changes, every score and stamp updates itself.
The full method ↗
