This study is designed to look at how snacking during exercise may help prevent low blood sugars while subjects are on the "closed loop artificial pancreas." This system uses a continuous glucose sensor, an insulin pump, and a computer program that automatically determines how much insulin to give based on the sensor glucose level.

This study will be a randomized cross-over study as subjects will be studied under both study conditions - Closed Loop (CL) alone vs. CL + proactive snacking protocol). Differences in the nadir blood glucose levels during exercise between the two study conditions will be the primary outcome measure.

Inclusion Criteria:

1. Age 13-45 years
2. Clinical diagnosis of T1D (formal antibody and/or genetic testing will not be required)
3. Duration of T1D ≥ 1 year
4. HbA1c ≤ 9 %
5. Treated with continuous subcutaneous insulin infusion (pump) for at least 3 months
6. Body weight > 40 kg (to accommodate phlebotomy)
7. Able to tolerate a 75-minute exercise period of moderate intensity
8. Be willing to wear 2 subcutaneous glucose sensors (placed under the skin like a pump site) simultaneously during the inpatient portions of this study.
9. Be in good general health without other acute or chronic illness that in the judgment of the investigator could interfere with the study or jeopardize subject safety
10. Normal hematocrit
11. Able to give consent (for children <18 years, permission from parents and subject assent will be required)
12. Female subjects of reproductive potential must be abstinent or consistently using appropriate family planning methods.

Exclusion Criteria:

1. Insulin resistant (defined as requiring > 1.5 units/kg/day at time of study enrollment)
2. Presence of any medical or psychiatric disorder that may interfere with subject safety or study conduct
3. Use of any medications (besides insulin) known to effect blood glucose levels, including oral or other systemic glucocorticoid therapy. Inhaled, intranasal, or rectal corticosteroid use is allowed along as not given within 2 weeks of the closed loop admissions. Use of topical glucocorticoids is allowable as long as affected skin area does not overlap with study device sites.
4. Subjects using herbal supplements will be excluded, due to the unknown effects of these supplements on glucose control
5. History of hypoglycemic seizure within last 3 months
6. Female subjects who are pregnant, lactating, or unwilling to be tested for pregnancy